feat: Add MCP UI Demo server and integrate UI rendering capabilities

feat: update Google and OpenAI SDKs with new features and fixes (#11395 )
* feat: update Google and OpenAI SDKs with new features and fixes - Updated Google SDK to ensure model paths are correctly formatted. - Enhanced OpenAI SDK to include support for image URLs in chat responses. - Added reasoning content handling in OpenAI chat responses and chunks. - Introduced Azure Anthropic provider configuration for Claude integration. * fix: azure error * fix: lint * fix: test * fix: test * fix type * fix comment * fix: redundant * chore resolution * fix: test * fix: comment * fix: comment * fix * feat: 添加 OpenRouter 推理中间件以支持内容过滤
2025-11-24 12:41:03 +08:00 · 2025-11-23 23:18:57 +08:00 · 2025-11-23 21:40:22 +08:00 · 2025-11-23 21:12:57 +08:00 · 2025-11-23 17:33:27 +08:00 · 2025-11-23 11:48:44 +08:00
205 changed files with 14704 additions and 2910 deletions
@@ -77,7 +77,7 @@ jobs:
        with:
          token: ${{ secrets.GITHUB_TOKEN }} # Use the built-in GITHUB_TOKEN for bot actions
          commit-message: "feat(bot): Weekly automated script run"
-          title: "🤖 Weekly Automated Update: ${{ env.CURRENT_DATE }}"
+          title: "🤖 Weekly Auto I18N Sync: ${{ env.CURRENT_DATE }}"
          body: |
            This PR includes changes generated by the weekly auto i18n.
            Review the changes before merging.
@@ -1,8 +1,8 @@
 diff --git a/dist/index.js b/dist/index.js
-index ff305b112779b718f21a636a27b1196125a332d9..cf32ff5086d4d9e56f8fe90c98724559083bafc3 100644
+index dc7b74ba55337c491cdf1ab3e39ca68cc4187884..ace8c90591288e42c2957e93c9bf7984f1b22444 100644
 --- a/dist/index.js
 +++ b/dist/index.js
-@@ -471,7 +471,7 @@ function convertToGoogleGenerativeAIMessages(prompt, options) {
+@@ -472,7 +472,7 @@ function convertToGoogleGenerativeAIMessages(prompt, options) {
 
 // src/get-model-path.ts
 function getModelPath(modelId) {
@@ -12,10 +12,10 @@ index ff305b112779b718f21a636a27b1196125a332d9..cf32ff5086d4d9e56f8fe90c98724559
 
 // src/google-generative-ai-options.ts
 diff --git a/dist/index.mjs b/dist/index.mjs
-index 57659290f1cec74878a385626ad75b2a4d5cd3fc..d04e5927ec3725b6ffdb80868bfa1b5a48849537 100644
+index 8390439c38cb7eaeb52080862cd6f4c58509e67c..a7647f2e11700dff7e1c8d4ae8f99d3637010733 100644
 --- a/dist/index.mjs
 +++ b/dist/index.mjs
-@@ -477,7 +477,7 @@ function convertToGoogleGenerativeAIMessages(prompt, options) {
+@@ -478,7 +478,7 @@ function convertToGoogleGenerativeAIMessages(prompt, options) {
 
 // src/get-model-path.ts
 function getModelPath(modelId) {
@@ -1,131 +0,0 @@
-diff --git a/dist/index.mjs b/dist/index.mjs
-index b3f018730a93639aad7c203f15fb1aeb766c73f4..ade2a43d66e9184799d072153df61ef7be4ea110 100644
--- a/dist/index.mjs
-+++ b/dist/index.mjs
-@@ -296,7 +296,14 @@ var HuggingFaceResponsesLanguageModel = class {
-       metadata: huggingfaceOptions == null ? void 0 : huggingfaceOptions.metadata,
-       instructions: huggingfaceOptions == null ? void 0 : huggingfaceOptions.instructions,
-       ...preparedTools && { tools: preparedTools },
-      ...preparedToolChoice && { tool_choice: preparedToolChoice }
-+      ...preparedToolChoice && { tool_choice: preparedToolChoice },
-+      ...(huggingfaceOptions?.reasoningEffort != null && {
-+        reasoning: {
-+          ...(huggingfaceOptions?.reasoningEffort != null && {
-+            effort: huggingfaceOptions.reasoningEffort,
-+          }),
-+        },
-+      }),
-     };
-     return { args: baseArgs, warnings };
-   }
-@@ -365,6 +372,20 @@ var HuggingFaceResponsesLanguageModel = class {
-           }
-           break;
-         }
-+        case 'reasoning': {
-+          for (const contentPart of part.content) {
-+            content.push({
-+              type: 'reasoning',
-+              text: contentPart.text,
-+              providerMetadata: {
-+                huggingface: {
-+                  itemId: part.id,
-+                },
-+              },
-+            });
-+          }
-+          break;
-+        }
-         case "mcp_call": {
-           content.push({
-             type: "tool-call",
-@@ -519,6 +540,11 @@ var HuggingFaceResponsesLanguageModel = class {
-                   id: value.item.call_id,
-                   toolName: value.item.name
-                 });
-+              } else if (value.item.type === 'reasoning') {
-+                controller.enqueue({
-+                  type: 'reasoning-start',
-+                  id: value.item.id,
-+                });
-               }
-               return;
-             }
-@@ -570,6 +596,22 @@ var HuggingFaceResponsesLanguageModel = class {
-               });
-               return;
-             }
-+            if (isReasoningDeltaChunk(value)) {
-+              controller.enqueue({
-+                type: 'reasoning-delta',
-+                id: value.item_id,
-+                delta: value.delta,
-+              });
-+              return;
-+            }
-+
-+            if (isReasoningEndChunk(value)) {
-+              controller.enqueue({
-+                type: 'reasoning-end',
-+                id: value.item_id,
-+              });
-+              return;
-+            }
-           },
-           flush(controller) {
-             controller.enqueue({
-@@ -593,7 +635,8 @@ var HuggingFaceResponsesLanguageModel = class {
- var huggingfaceResponsesProviderOptionsSchema = z2.object({
-   metadata: z2.record(z2.string(), z2.string()).optional(),
-   instructions: z2.string().optional(),
-  strictJsonSchema: z2.boolean().optional()
-+  strictJsonSchema: z2.boolean().optional(),
-+  reasoningEffort: z2.string().optional(),
- });
- var huggingfaceResponsesResponseSchema = z2.object({
-   id: z2.string(),
-@@ -727,12 +770,31 @@ var responseCreatedChunkSchema = z2.object({
-     model: z2.string()
-   })
- });
-+var reasoningTextDeltaChunkSchema = z2.object({
-+  type: z2.literal('response.reasoning_text.delta'),
-+  item_id: z2.string(),
-+  output_index: z2.number(),
-+  content_index: z2.number(),
-+  delta: z2.string(),
-+  sequence_number: z2.number(),
-+});
-+
-+var reasoningTextEndChunkSchema = z2.object({
-+  type: z2.literal('response.reasoning_text.done'),
-+  item_id: z2.string(),
-+  output_index: z2.number(),
-+  content_index: z2.number(),
-+  text: z2.string(),
-+  sequence_number: z2.number(),
-+});
- var huggingfaceResponsesChunkSchema = z2.union([
-   responseOutputItemAddedSchema,
-   responseOutputItemDoneSchema,
-   textDeltaChunkSchema,
-   responseCompletedChunkSchema,
-   responseCreatedChunkSchema,
-+  reasoningTextDeltaChunkSchema,
-+  reasoningTextEndChunkSchema,
-   z2.object({ type: z2.string() }).loose()
-   // fallback for unknown chunks
- ]);
-@@ -751,6 +813,12 @@ function isResponseCompletedChunk(chunk) {
- function isResponseCreatedChunk(chunk) {
-   return chunk.type === "response.created";
- }
-+function isReasoningDeltaChunk(chunk) {
-+  return chunk.type === 'response.reasoning_text.delta';
-+}
-+function isReasoningEndChunk(chunk) {
-+  return chunk.type === 'response.reasoning_text.done';
-+}
- 
- // src/huggingface-provider.ts
- function createHuggingFace(options = {}) {
@@ -0,0 +1,140 @@
+diff --git a/dist/index.js b/dist/index.js
+index 73045a7d38faafdc7f7d2cd79d7ff0e2b031056b..8d948c9ac4ea4b474db9ef3c5491961e7fcf9a07 100644
+--- a/dist/index.js
+++ b/dist/index.js
+@@ -421,6 +421,17 @@ var OpenAICompatibleChatLanguageModel = class {
+         text: reasoning
+       });
+     }
+    if (choice.message.images) {
+      for (const image of choice.message.images) {
+        const match1 = image.image_url.url.match(/^data:([^;]+)/)
+        const match2 = image.image_url.url.match(/^data:[^;]*;base64,(.+)$/);
+        content.push({
+          type: 'file',
+          mediaType: match1 ? (match1[1] ?? 'image/jpeg') : 'image/jpeg',
+          data: match2 ? match2[1] : image.image_url.url,
+        });
+      }
+    }
+     if (choice.message.tool_calls != null) {
+       for (const toolCall of choice.message.tool_calls) {
+         content.push({
+@@ -598,6 +609,17 @@ var OpenAICompatibleChatLanguageModel = class {
+                 delta: delta.content
+               });
+             }
+            if (delta.images) {
+              for (const image of delta.images) {
+                const match1 = image.image_url.url.match(/^data:([^;]+)/)
+                const match2 = image.image_url.url.match(/^data:[^;]*;base64,(.+)$/);
+                controller.enqueue({
+                  type: 'file',
+                  mediaType: match1 ? (match1[1] ?? 'image/jpeg') : 'image/jpeg',
+                  data: match2 ? match2[1] : image.image_url.url,
+                });
+              }
+            }
+             if (delta.tool_calls != null) {
+               for (const toolCallDelta of delta.tool_calls) {
+                 const index = toolCallDelta.index;
+@@ -765,6 +787,14 @@ var OpenAICompatibleChatResponseSchema = import_v43.z.object({
+               arguments: import_v43.z.string()
+             })
+           })
+        ).nullish(),
+        images: import_v43.z.array(
+          import_v43.z.object({
+            type: import_v43.z.literal('image_url'),
+            image_url: import_v43.z.object({
+              url: import_v43.z.string(),
+            })
+          })
+         ).nullish()
+       }),
+       finish_reason: import_v43.z.string().nullish()
+@@ -795,6 +825,14 @@ var createOpenAICompatibleChatChunkSchema = (errorSchema) => import_v43.z.union(
+                 arguments: import_v43.z.string().nullish()
+               })
+             })
+          ).nullish(),
+          images: import_v43.z.array(
+            import_v43.z.object({
+              type: import_v43.z.literal('image_url'),
+              image_url: import_v43.z.object({
+                url: import_v43.z.string(),
+              })
+            })
+           ).nullish()
+         }).nullish(),
+         finish_reason: import_v43.z.string().nullish()
+diff --git a/dist/index.mjs b/dist/index.mjs
+index 1c2b9560bbfbfe10cb01af080aeeed4ff59db29c..2c8ddc4fc9bfc5e7e06cfca105d197a08864c427 100644
+--- a/dist/index.mjs
+++ b/dist/index.mjs
+@@ -405,6 +405,17 @@ var OpenAICompatibleChatLanguageModel = class {
+         text: reasoning
+       });
+     }
+    if (choice.message.images) {
+      for (const image of choice.message.images) {
+        const match1 = image.image_url.url.match(/^data:([^;]+)/)
+        const match2 = image.image_url.url.match(/^data:[^;]*;base64,(.+)$/);
+        content.push({
+          type: 'file',
+          mediaType: match1 ? (match1[1] ?? 'image/jpeg') : 'image/jpeg',
+          data: match2 ? match2[1] : image.image_url.url,
+        });
+      }
+    }
+     if (choice.message.tool_calls != null) {
+       for (const toolCall of choice.message.tool_calls) {
+         content.push({
+@@ -582,6 +593,17 @@ var OpenAICompatibleChatLanguageModel = class {
+                 delta: delta.content
+               });
+             }
+            if (delta.images) {
+              for (const image of delta.images) {
+                const match1 = image.image_url.url.match(/^data:([^;]+)/)
+                const match2 = image.image_url.url.match(/^data:[^;]*;base64,(.+)$/);
+                controller.enqueue({
+                  type: 'file',
+                  mediaType: match1 ? (match1[1] ?? 'image/jpeg') : 'image/jpeg',
+                  data: match2 ? match2[1] : image.image_url.url,
+                });
+              }
+            }
+             if (delta.tool_calls != null) {
+               for (const toolCallDelta of delta.tool_calls) {
+                 const index = toolCallDelta.index;
+@@ -749,6 +771,14 @@ var OpenAICompatibleChatResponseSchema = z3.object({
+               arguments: z3.string()
+             })
+           })
+        ).nullish(),
+        images: z3.array(
+          z3.object({
+            type: z3.literal('image_url'),
+            image_url: z3.object({
+              url: z3.string(),
+            })
+          })
+         ).nullish()
+       }),
+       finish_reason: z3.string().nullish()
+@@ -779,6 +809,14 @@ var createOpenAICompatibleChatChunkSchema = (errorSchema) => z3.union([
+                 arguments: z3.string().nullish()
+               })
+             })
+          ).nullish(),
+          images: z3.array(
+            z3.object({
+              type: z3.literal('image_url'),
+              image_url: z3.object({
+                url: z3.string(),
+              })
+            })
+           ).nullish()
+         }).nullish(),
+         finish_reason: z3.string().nullish()
@@ -1,5 +1,5 @@
 diff --git a/dist/index.js b/dist/index.js
-index 992c85ac6656e51c3471af741583533c5a7bf79f..83c05952a07aebb95fc6c62f9ddb8aa96b52ac0d 100644
+index 7481f3b3511078068d87d03855b568b20bb86971..8ac5ec28d2f7ad1b3b0d3f8da945c75674e59637 100644
 --- a/dist/index.js
 +++ b/dist/index.js
@@ -274,6 +274,7 @@ var openaiChatResponseSchema = (0, import_provider_utils3.lazyValidator)(
@@ -18,7 +18,7 @@ index 992c85ac6656e51c3471af741583533c5a7bf79f..83c05952a07aebb95fc6c62f9ddb8aa9
               tool_calls: import_v42.z.array(
                 import_v42.z.object({
                   index: import_v42.z.number(),
-@@ -785,6 +787,13 @@ var OpenAIChatLanguageModel = class {
+@@ -795,6 +797,13 @@ var OpenAIChatLanguageModel = class {
     if (text != null && text.length > 0) {
       content.push({ type: "text", text });
     }
@@ -32,7 +32,7 @@ index 992c85ac6656e51c3471af741583533c5a7bf79f..83c05952a07aebb95fc6c62f9ddb8aa9
     for (const toolCall of (_a = choice.message.tool_calls) != null ? _a : []) {
       content.push({
         type: "tool-call",
-@@ -866,6 +875,7 @@ var OpenAIChatLanguageModel = class {
+@@ -876,6 +885,7 @@ var OpenAIChatLanguageModel = class {
     };
     let metadataExtracted = false;
     let isActiveText = false;
@@ -40,7 +40,7 @@ index 992c85ac6656e51c3471af741583533c5a7bf79f..83c05952a07aebb95fc6c62f9ddb8aa9
     const providerMetadata = { openai: {} };
     return {
       stream: response.pipeThrough(
-@@ -923,6 +933,21 @@ var OpenAIChatLanguageModel = class {
+@@ -933,6 +943,21 @@ var OpenAIChatLanguageModel = class {
               return;
             }
             const delta = choice.delta;
@@ -62,7 +62,7 @@ index 992c85ac6656e51c3471af741583533c5a7bf79f..83c05952a07aebb95fc6c62f9ddb8aa9
             if (delta.content != null) {
               if (!isActiveText) {
                 controller.enqueue({ type: "text-start", id: "0" });
-@@ -1035,6 +1060,9 @@ var OpenAIChatLanguageModel = class {
+@@ -1045,6 +1070,9 @@ var OpenAIChatLanguageModel = class {
             }
           },
           flush(controller) {
@@ -14,7 +14,7 @@
      }
    },
    "enabled": true,
-    "includes": ["**/*.json", "!*.json", "!**/package.json"]
+    "includes": ["**/*.json", "!*.json", "!**/package.json", "!coverage/**"]
  },
  "css": {
    "formatter": {
@@ -23,7 +23,7 @@
  },
  "files": {
    "ignoreUnknown": false,
-    "includes": ["**", "!**/.claude/**"],
+    "includes": ["**", "!**/.claude/**", "!**/.vscode/**"],
    "maxSize": 2097152
  },
  "formatter": {
@@ -1,6 +1,6 @@
 {
  "name": "CherryStudio",
-  "version": "1.6.5",
+  "version": "1.7.0-rc.1",
  "private": true,
  "description": "A powerful AI assistant for producer.",
  "main": "./out/main/index.js",
@@ -74,9 +74,10 @@
    "format:check": "biome format && biome lint",
    "prepare": "git config blame.ignoreRevsFile .git-blame-ignore-revs && husky",
    "claude": "dotenv -e .env -- claude",
-    "release:aicore:alpha": "yarn workspace @cherrystudio/ai-core version prerelease --immediate && yarn workspace @cherrystudio/ai-core npm publish --tag alpha --access public",
-    "release:aicore:beta": "yarn workspace @cherrystudio/ai-core version prerelease --immediate && yarn workspace @cherrystudio/ai-core npm publish --tag beta --access public",
-    "release:aicore": "yarn workspace @cherrystudio/ai-core version patch --immediate && yarn workspace @cherrystudio/ai-core npm publish --access public"
+    "release:aicore:alpha": "yarn workspace @cherrystudio/ai-core version prerelease --preid alpha --immediate && yarn workspace @cherrystudio/ai-core build && yarn workspace @cherrystudio/ai-core npm publish --tag alpha --access public",
+    "release:aicore:beta": "yarn workspace @cherrystudio/ai-core version prerelease --preid beta --immediate && yarn workspace @cherrystudio/ai-core build && yarn workspace @cherrystudio/ai-core npm publish --tag beta --access public",
+    "release:aicore": "yarn workspace @cherrystudio/ai-core version patch --immediate && yarn workspace @cherrystudio/ai-core build && yarn workspace @cherrystudio/ai-core npm publish --access public",
+    "release:ai-sdk-provider": "yarn workspace @cherrystudio/ai-sdk-provider version patch --immediate && yarn workspace @cherrystudio/ai-sdk-provider build && yarn workspace @cherrystudio/ai-sdk-provider npm publish --access public"
  },
  "dependencies": {
    "@anthropic-ai/claude-agent-sdk": "patch:@anthropic-ai/claude-agent-sdk@npm%3A0.1.30#~/.yarn/patches/@anthropic-ai-claude-agent-sdk-npm-0.1.30-b50a299674.patch",
@@ -85,6 +86,7 @@
    "@napi-rs/system-ocr": "patch:@napi-rs/system-ocr@npm%3A1.0.2#~/.yarn/patches/@napi-rs-system-ocr-npm-1.0.2-59e7a78e8b.patch",
    "@paymoapp/electron-shutdown-handler": "^1.1.2",
    "@strongtz/win32-arm64-msvc": "^0.4.7",
+    "emoji-picker-element-data": "^1",
    "express": "^5.1.0",
    "font-list": "^2.0.0",
    "graceful-fs": "^4.2.11",
@@ -107,13 +109,17 @@
    "@agentic/exa": "^7.3.3",
    "@agentic/searxng": "^7.3.3",
    "@agentic/tavily": "^7.3.3",
-    "@ai-sdk/amazon-bedrock": "^3.0.53",
+    "@ai-sdk/amazon-bedrock": "^3.0.56",
+    "@ai-sdk/anthropic": "^2.0.45",
    "@ai-sdk/cerebras": "^1.0.31",
-    "@ai-sdk/gateway": "^2.0.9",
-    "@ai-sdk/google-vertex": "^3.0.62",
-    "@ai-sdk/huggingface": "patch:@ai-sdk/huggingface@npm%3A0.0.8#~/.yarn/patches/@ai-sdk-huggingface-npm-0.0.8-d4d0aaac93.patch",
-    "@ai-sdk/mistral": "^2.0.23",
-    "@ai-sdk/perplexity": "^2.0.17",
+    "@ai-sdk/gateway": "^2.0.13",
+    "@ai-sdk/google": "patch:@ai-sdk/google@npm%3A2.0.40#~/.yarn/patches/@ai-sdk-google-npm-2.0.40-47e0eeee83.patch",
+    "@ai-sdk/google-vertex": "^3.0.72",
+    "@ai-sdk/huggingface": "^0.0.10",
+    "@ai-sdk/mistral": "^2.0.24",
+    "@ai-sdk/openai": "patch:@ai-sdk/openai@npm%3A2.0.71#~/.yarn/patches/@ai-sdk-openai-npm-2.0.71-a88ef00525.patch",
+    "@ai-sdk/perplexity": "^2.0.20",
+    "@ai-sdk/test-server": "^0.0.1",
    "@ant-design/v5-patch-for-react-19": "^1.0.3",
    "@anthropic-ai/sdk": "^0.41.0",
    "@anthropic-ai/vertex-sdk": "patch:@anthropic-ai/vertex-sdk@npm%3A0.11.4#~/.yarn/patches/@anthropic-ai-vertex-sdk-npm-0.11.4-c19cb41edb.patch",
@@ -121,7 +127,7 @@
    "@aws-sdk/client-bedrock-runtime": "^3.910.0",
    "@aws-sdk/client-s3": "^3.910.0",
    "@biomejs/biome": "2.2.4",
-    "@cherrystudio/ai-core": "workspace:^1.0.0-alpha.18",
+    "@cherrystudio/ai-core": "workspace:^1.0.9",
    "@cherrystudio/embedjs": "^0.1.31",
    "@cherrystudio/embedjs-libsql": "^0.1.31",
    "@cherrystudio/embedjs-loader-csv": "^0.1.31",
@@ -135,7 +141,7 @@
    "@cherrystudio/embedjs-ollama": "^0.1.31",
    "@cherrystudio/embedjs-openai": "^0.1.31",
    "@cherrystudio/extension-table-plus": "workspace:^",
-    "@cherrystudio/openai": "^6.5.0",
+    "@cherrystudio/openai": "^6.9.0",
    "@dnd-kit/core": "^6.3.1",
    "@dnd-kit/modifiers": "^9.0.0",
    "@dnd-kit/sortable": "^10.0.0",
@@ -154,18 +160,19 @@
    "@langchain/community": "^1.0.0",
    "@langchain/core": "patch:@langchain/core@npm%3A1.0.2#~/.yarn/patches/@langchain-core-npm-1.0.2-183ef83fe4.patch",
    "@langchain/openai": "patch:@langchain/openai@npm%3A1.0.0#~/.yarn/patches/@langchain-openai-npm-1.0.0-474d0ad9d4.patch",
+    "@mcp-ui/client": "^5.14.1",
    "@mistralai/mistralai": "^1.7.5",
    "@modelcontextprotocol/sdk": "^1.17.5",
    "@mozilla/readability": "^0.6.0",
    "@notionhq/client": "^2.2.15",
-    "@openrouter/ai-sdk-provider": "^1.2.0",
+    "@openrouter/ai-sdk-provider": "^1.2.5",
    "@opentelemetry/api": "^1.9.0",
    "@opentelemetry/core": "2.0.0",
    "@opentelemetry/exporter-trace-otlp-http": "^0.200.0",
    "@opentelemetry/sdk-trace-base": "^2.0.0",
    "@opentelemetry/sdk-trace-node": "^2.0.0",
    "@opentelemetry/sdk-trace-web": "^2.0.0",
-    "@opeoginni/github-copilot-openai-compatible": "0.1.19",
+    "@opeoginni/github-copilot-openai-compatible": "0.1.21",
    "@playwright/test": "^1.52.0",
    "@radix-ui/react-context-menu": "^2.2.16",
    "@reduxjs/toolkit": "^2.2.5",
@@ -234,7 +241,7 @@
    "@viz-js/lang-dot": "^1.0.5",
    "@viz-js/viz": "^3.14.0",
    "@xyflow/react": "^12.4.4",
-    "ai": "^5.0.90",
+    "ai": "^5.0.98",
    "antd": "patch:antd@npm%3A5.27.0#~/.yarn/patches/antd-npm-5.27.0-aa91c36546.patch",
    "archiver": "^7.0.1",
    "async-mutex": "^0.5.0",
@@ -407,8 +414,11 @@
    "@langchain/openai@npm:^0.3.16": "patch:@langchain/openai@npm%3A1.0.0#~/.yarn/patches/@langchain-openai-npm-1.0.0-474d0ad9d4.patch",
    "@langchain/openai@npm:>=0.2.0 <0.7.0": "patch:@langchain/openai@npm%3A1.0.0#~/.yarn/patches/@langchain-openai-npm-1.0.0-474d0ad9d4.patch",
    "@ai-sdk/openai@npm:2.0.64": "patch:@ai-sdk/openai@npm%3A2.0.64#~/.yarn/patches/@ai-sdk-openai-npm-2.0.64-48f99f5bf3.patch",
-    "@ai-sdk/openai@npm:^2.0.42": "patch:@ai-sdk/openai@npm%3A2.0.64#~/.yarn/patches/@ai-sdk-openai-npm-2.0.64-48f99f5bf3.patch",
-    "@ai-sdk/google@npm:2.0.31": "patch:@ai-sdk/google@npm%3A2.0.31#~/.yarn/patches/@ai-sdk-google-npm-2.0.31-b0de047210.patch"
+    "@ai-sdk/openai@npm:^2.0.42": "patch:@ai-sdk/openai@npm%3A2.0.71#~/.yarn/patches/@ai-sdk-openai-npm-2.0.71-a88ef00525.patch",
+    "@ai-sdk/google@npm:2.0.40": "patch:@ai-sdk/google@npm%3A2.0.40#~/.yarn/patches/@ai-sdk-google-npm-2.0.40-47e0eeee83.patch",
+    "@ai-sdk/openai@npm:2.0.71": "patch:@ai-sdk/openai@npm%3A2.0.71#~/.yarn/patches/@ai-sdk-openai-npm-2.0.71-a88ef00525.patch",
+    "@ai-sdk/openai-compatible@npm:1.0.27": "patch:@ai-sdk/openai-compatible@npm%3A1.0.27#~/.yarn/patches/@ai-sdk-openai-compatible-npm-1.0.27-06f74278cf.patch",
+    "@ai-sdk/openai-compatible@npm:^1.0.19": "patch:@ai-sdk/openai-compatible@npm%3A1.0.27#~/.yarn/patches/@ai-sdk-openai-compatible-npm-1.0.27-06f74278cf.patch"
  },
  "packageManager": "yarn@4.9.1",
  "lint-staged": {
@@ -1,6 +1,6 @@
 {
  "name": "@cherrystudio/ai-sdk-provider",
-  "version": "0.1.0",
+  "version": "0.1.3",
  "description": "Cherry Studio AI SDK provider bundle with CherryIN routing.",
  "keywords": [
    "ai-sdk",
@@ -42,7 +42,7 @@
  },
  "dependencies": {
    "@ai-sdk/provider": "^2.0.0",
-    "@ai-sdk/provider-utils": "^3.0.12"
+    "@ai-sdk/provider-utils": "^3.0.17"
  },
  "devDependencies": {
    "tsdown": "^0.13.3",
@@ -67,6 +67,10 @@ export interface CherryInProviderSettings {
   * Optional static headers applied to every request.
   */
  headers?: HeadersInput
+  /**
+   * Optional endpoint type to distinguish different endpoint behaviors.
+   */
+  endpointType?: 'openai' | 'openai-response' | 'anthropic' | 'gemini' | 'image-generation' | 'jina-rerank'
 }

 export interface CherryInProvider extends ProviderV2 {
@@ -151,7 +155,8 @@ export const createCherryIn = (options: CherryInProviderSettings = {}): CherryIn
    baseURL = DEFAULT_CHERRYIN_BASE_URL,
    anthropicBaseURL = DEFAULT_CHERRYIN_ANTHROPIC_BASE_URL,
    geminiBaseURL = DEFAULT_CHERRYIN_GEMINI_BASE_URL,
-    fetch
+    fetch,
+    endpointType
  } = options

  const getJsonHeaders = createJsonHeadersGetter(options)
@@ -205,7 +210,7 @@ export const createCherryIn = (options: CherryInProviderSettings = {}): CherryIn
      fetch
    })

-  const createChatModel = (modelId: string, settings: OpenAIProviderSettings = {}) => {
+  const createChatModelByModelId = (modelId: string, settings: OpenAIProviderSettings = {}) => {
    if (isAnthropicModel(modelId)) {
      return createAnthropicModel(modelId)
    }
@@ -223,6 +228,29 @@ export const createCherryIn = (options: CherryInProviderSettings = {}): CherryIn
    })
  }

+  const createChatModel = (modelId: string, settings: OpenAIProviderSettings = {}) => {
+    if (!endpointType) return createChatModelByModelId(modelId, settings)
+    switch (endpointType) {
+      case 'anthropic':
+        return createAnthropicModel(modelId)
+      case 'gemini':
+        return createGeminiModel(modelId)
+      case 'openai':
+        return createOpenAIChatModel(modelId)
+      case 'openai-response':
+      default:
+        return new OpenAIResponsesLanguageModel(modelId, {
+          provider: `${CHERRYIN_PROVIDER_NAME}.openai`,
+          url,
+          headers: () => ({
+            ...getJsonHeaders(),
+            ...settings.headers
+          }),
+          fetch
+        })
+    }
+  }
+
  const createCompletionModel = (modelId: string, settings: OpenAIProviderSettings = {}) =>
    new OpenAICompletionLanguageModel(modelId, {
      provider: `${CHERRYIN_PROVIDER_NAME}.completion`,
@@ -71,7 +71,7 @@ Cherry Studio AI Core 是一个基于 Vercel AI SDK 的统一 AI Provider 接口
 ## 安装

 ```bash
-npm install @cherrystudio/ai-core ai
+npm install @cherrystudio/ai-core ai @ai-sdk/google @ai-sdk/openai
 ```

 ### React Native
@@ -1,6 +1,6 @@
 {
  "name": "@cherrystudio/ai-core",
-  "version": "1.0.1",
+  "version": "1.0.9",
  "description": "Cherry Studio AI Core - Unified AI Provider Interface Based on Vercel AI SDK",
  "main": "dist/index.js",
  "module": "dist/index.mjs",
@@ -33,19 +33,19 @@
  },
  "homepage": "https://github.com/CherryHQ/cherry-studio#readme",
  "peerDependencies": {
+    "@ai-sdk/google": "^2.0.36",
+    "@ai-sdk/openai": "^2.0.64",
+    "@cherrystudio/ai-sdk-provider": "^0.1.3",
    "ai": "^5.0.26"
  },
  "dependencies": {
-    "@ai-sdk/anthropic": "^2.0.43",
-    "@ai-sdk/azure": "^2.0.66",
-    "@ai-sdk/deepseek": "^1.0.27",
-    "@ai-sdk/google": "patch:@ai-sdk/google@npm%3A2.0.31#~/.yarn/patches/@ai-sdk-google-npm-2.0.31-b0de047210.patch",
-    "@ai-sdk/openai": "patch:@ai-sdk/openai@npm%3A2.0.64#~/.yarn/patches/@ai-sdk-openai-npm-2.0.64-48f99f5bf3.patch",
-    "@ai-sdk/openai-compatible": "^1.0.26",
+    "@ai-sdk/anthropic": "^2.0.45",
+    "@ai-sdk/azure": "^2.0.73",
+    "@ai-sdk/deepseek": "^1.0.29",
+    "@ai-sdk/openai-compatible": "patch:@ai-sdk/openai-compatible@npm%3A1.0.27#~/.yarn/patches/@ai-sdk-openai-compatible-npm-1.0.27-06f74278cf.patch",
    "@ai-sdk/provider": "^2.0.0",
-    "@ai-sdk/provider-utils": "^3.0.16",
-    "@ai-sdk/xai": "^2.0.31",
-    "@cherrystudio/ai-sdk-provider": "workspace:*",
+    "@ai-sdk/provider-utils": "^3.0.17",
+    "@ai-sdk/xai": "^2.0.34",
    "zod": "^4.1.5"
  },
  "devDependencies": {
@@ -0,0 +1,180 @@
+/**
+ * Mock Provider Instances
+ * Provides mock implementations for all supported AI providers
+ */
+
+import type { ImageModelV2, LanguageModelV2 } from '@ai-sdk/provider'
+import { vi } from 'vitest'
+
+/**
+ * Creates a mock language model with customizable behavior
+ */
+export function createMockLanguageModel(overrides?: Partial<LanguageModelV2>): LanguageModelV2 {
+  return {
+    specificationVersion: 'v1',
+    provider: 'mock-provider',
+    modelId: 'mock-model',
+    defaultObjectGenerationMode: 'tool',
+
+    doGenerate: vi.fn().mockResolvedValue({
+      text: 'Mock response text',
+      finishReason: 'stop',
+      usage: {
+        promptTokens: 10,
+        completionTokens: 20,
+        totalTokens: 30
+      },
+      rawCall: { rawPrompt: null, rawSettings: {} },
+      rawResponse: { headers: {} },
+      warnings: []
+    }),
+
+    doStream: vi.fn().mockReturnValue({
+      stream: (async function* () {
+        yield {
+          type: 'text-delta',
+          textDelta: 'Mock '
+        }
+        yield {
+          type: 'text-delta',
+          textDelta: 'streaming '
+        }
+        yield {
+          type: 'text-delta',
+          textDelta: 'response'
+        }
+        yield {
+          type: 'finish',
+          finishReason: 'stop',
+          usage: {
+            promptTokens: 10,
+            completionTokens: 15,
+            totalTokens: 25
+          }
+        }
+      })(),
+      rawCall: { rawPrompt: null, rawSettings: {} },
+      rawResponse: { headers: {} },
+      warnings: []
+    }),
+
+    ...overrides
+  } as LanguageModelV2
+}
+
+/**
+ * Creates a mock image model with customizable behavior
+ */
+export function createMockImageModel(overrides?: Partial<ImageModelV2>): ImageModelV2 {
+  return {
+    specificationVersion: 'v2',
+    provider: 'mock-provider',
+    modelId: 'mock-image-model',
+
+    doGenerate: vi.fn().mockResolvedValue({
+      images: [
+        {
+          base64: 'mock-base64-image-data',
+          uint8Array: new Uint8Array([1, 2, 3, 4, 5]),
+          mimeType: 'image/png'
+        }
+      ],
+      warnings: []
+    }),
+
+    ...overrides
+  } as ImageModelV2
+}
+
+/**
+ * Mock provider configurations for testing
+ */
+export const mockProviderConfigs = {
+  openai: {
+    apiKey: 'sk-test-openai-key-123456789',
+    baseURL: 'https://api.openai.com/v1',
+    organization: 'test-org'
+  },
+
+  anthropic: {
+    apiKey: 'sk-ant-test-key-123456789',
+    baseURL: 'https://api.anthropic.com'
+  },
+
+  google: {
+    apiKey: 'test-google-api-key-123456789',
+    baseURL: 'https://generativelanguage.googleapis.com/v1'
+  },
+
+  xai: {
+    apiKey: 'xai-test-key-123456789',
+    baseURL: 'https://api.x.ai/v1'
+  },
+
+  azure: {
+    apiKey: 'test-azure-key-123456789',
+    resourceName: 'test-resource',
+    deployment: 'test-deployment'
+  },
+
+  deepseek: {
+    apiKey: 'sk-test-deepseek-key-123456789',
+    baseURL: 'https://api.deepseek.com/v1'
+  },
+
+  openrouter: {
+    apiKey: 'sk-or-test-key-123456789',
+    baseURL: 'https://openrouter.ai/api/v1'
+  },
+
+  huggingface: {
+    apiKey: 'hf_test_key_123456789',
+    baseURL: 'https://api-inference.huggingface.co'
+  },
+
+  'openai-compatible': {
+    apiKey: 'test-compatible-key-123456789',
+    baseURL: 'https://api.example.com/v1',
+    name: 'test-provider'
+  },
+
+  'openai-chat': {
+    apiKey: 'sk-test-chat-key-123456789',
+    baseURL: 'https://api.openai.com/v1'
+  }
+} as const
+
+/**
+ * Mock provider instances for testing
+ */
+export const mockProviderInstances = {
+  openai: {
+    name: 'openai-mock',
+    languageModel: createMockLanguageModel({ provider: 'openai', modelId: 'gpt-4' }),
+    imageModel: createMockImageModel({ provider: 'openai', modelId: 'dall-e-3' })
+  },
+
+  anthropic: {
+    name: 'anthropic-mock',
+    languageModel: createMockLanguageModel({ provider: 'anthropic', modelId: 'claude-3-5-sonnet-20241022' })
+  },
+
+  google: {
+    name: 'google-mock',
+    languageModel: createMockLanguageModel({ provider: 'google', modelId: 'gemini-2.0-flash-exp' }),
+    imageModel: createMockImageModel({ provider: 'google', modelId: 'imagen-3.0-generate-001' })
+  },
+
+  xai: {
+    name: 'xai-mock',
+    languageModel: createMockLanguageModel({ provider: 'xai', modelId: 'grok-2-latest' }),
+    imageModel: createMockImageModel({ provider: 'xai', modelId: 'grok-2-image-latest' })
+  },
+
+  deepseek: {
+    name: 'deepseek-mock',
+    languageModel: createMockLanguageModel({ provider: 'deepseek', modelId: 'deepseek-chat' })
+  }
+}
+
+export type ProviderId = keyof typeof mockProviderConfigs
@@ -0,0 +1,331 @@
+/**
+ * Mock Responses
+ * Provides realistic mock responses for all provider types
+ */
+
+import { jsonSchema, type ModelMessage, type Tool } from 'ai'
+
+/**
+ * Standard test messages for all scenarios
+ */
+export const testMessages = {
+  simple: [{ role: 'user' as const, content: 'Hello, how are you?' }],
+
+  conversation: [
+    { role: 'user' as const, content: 'What is the capital of France?' },
+    { role: 'assistant' as const, content: 'The capital of France is Paris.' },
+    { role: 'user' as const, content: 'What is its population?' }
+  ],
+
+  withSystem: [
+    { role: 'system' as const, content: 'You are a helpful assistant that provides concise answers.' },
+    { role: 'user' as const, content: 'Explain quantum computing in one sentence.' }
+  ],
+
+  withImages: [
+    {
+      role: 'user' as const,
+      content: [
+        { type: 'text' as const, text: 'What is in this image?' },
+        {
+          type: 'image' as const,
+          image:
+            'data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mNk+M9QDwADhgGAWjR9awAAAABJRU5ErkJggg=='
+        }
+      ]
+    }
+  ],
+
+  toolUse: [{ role: 'user' as const, content: 'What is the weather in San Francisco?' }],
+
+  multiTurn: [
+    { role: 'user' as const, content: 'Can you help me with a math problem?' },
+    { role: 'assistant' as const, content: 'Of course! What math problem would you like help with?' },
+    { role: 'user' as const, content: 'What is 15 * 23?' },
+    { role: 'assistant' as const, content: '15 * 23 = 345' },
+    { role: 'user' as const, content: 'Now divide that by 5' }
+  ]
+} satisfies Record<string, ModelMessage[]>
+
+/**
+ * Standard test tools for tool calling scenarios
+ */
+export const testTools: Record<string, Tool> = {
+  getWeather: {
+    description: 'Get the current weather in a given location',
+    inputSchema: jsonSchema({
+      type: 'object',
+      properties: {
+        location: {
+          type: 'string',
+          description: 'The city and state, e.g. San Francisco, CA'
+        },
+        unit: {
+          type: 'string',
+          enum: ['celsius', 'fahrenheit'],
+          description: 'The temperature unit to use'
+        }
+      },
+      required: ['location']
+    }),
+    execute: async ({ location, unit = 'fahrenheit' }) => {
+      return {
+        location,
+        temperature: unit === 'celsius' ? 22 : 72,
+        unit,
+        condition: 'sunny'
+      }
+    }
+  },
+
+  calculate: {
+    description: 'Perform a mathematical calculation',
+    inputSchema: jsonSchema({
+      type: 'object',
+      properties: {
+        operation: {
+          type: 'string',
+          enum: ['add', 'subtract', 'multiply', 'divide'],
+          description: 'The operation to perform'
+        },
+        a: {
+          type: 'number',
+          description: 'The first number'
+        },
+        b: {
+          type: 'number',
+          description: 'The second number'
+        }
+      },
+      required: ['operation', 'a', 'b']
+    }),
+    execute: async ({ operation, a, b }) => {
+      const operations = {
+        add: (x: number, y: number) => x + y,
+        subtract: (x: number, y: number) => x - y,
+        multiply: (x: number, y: number) => x * y,
+        divide: (x: number, y: number) => x / y
+      }
+      return { result: operations[operation as keyof typeof operations](a, b) }
+    }
+  },
+
+  searchDatabase: {
+    description: 'Search for information in a database',
+    inputSchema: jsonSchema({
+      type: 'object',
+      properties: {
+        query: {
+          type: 'string',
+          description: 'The search query'
+        },
+        limit: {
+          type: 'number',
+          description: 'Maximum number of results to return',
+          default: 10
+        }
+      },
+      required: ['query']
+    }),
+    execute: async ({ query, limit = 10 }) => {
+      return {
+        results: [
+          { id: 1, title: `Result 1 for ${query}`, relevance: 0.95 },
+          { id: 2, title: `Result 2 for ${query}`, relevance: 0.87 }
+        ].slice(0, limit)
+      }
+    }
+  }
+}
+
+/**
+ * Mock streaming chunks for different providers
+ */
+export const mockStreamingChunks = {
+  text: [
+    { type: 'text-delta' as const, textDelta: 'Hello' },
+    { type: 'text-delta' as const, textDelta: ', ' },
+    { type: 'text-delta' as const, textDelta: 'this ' },
+    { type: 'text-delta' as const, textDelta: 'is ' },
+    { type: 'text-delta' as const, textDelta: 'a ' },
+    { type: 'text-delta' as const, textDelta: 'test.' }
+  ],
+
+  withToolCall: [
+    { type: 'text-delta' as const, textDelta: 'Let me check the weather for you.' },
+    {
+      type: 'tool-call-delta' as const,
+      toolCallType: 'function' as const,
+      toolCallId: 'call_123',
+      toolName: 'getWeather',
+      argsTextDelta: '{"location":'
+    },
+    {
+      type: 'tool-call-delta' as const,
+      toolCallType: 'function' as const,
+      toolCallId: 'call_123',
+      toolName: 'getWeather',
+      argsTextDelta: ' "San Francisco, CA"}'
+    },
+    {
+      type: 'tool-call' as const,
+      toolCallType: 'function' as const,
+      toolCallId: 'call_123',
+      toolName: 'getWeather',
+      args: { location: 'San Francisco, CA' }
+    }
+  ],
+
+  withFinish: [
+    { type: 'text-delta' as const, textDelta: 'Complete response.' },
+    {
+      type: 'finish' as const,
+      finishReason: 'stop' as const,
+      usage: {
+        promptTokens: 10,
+        completionTokens: 5,
+        totalTokens: 15
+      }
+    }
+  ]
+}
+
+/**
+ * Mock complete responses for non-streaming scenarios
+ */
+export const mockCompleteResponses = {
+  simple: {
+    text: 'This is a simple response.',
+    finishReason: 'stop' as const,
+    usage: {
+      promptTokens: 15,
+      completionTokens: 8,
+      totalTokens: 23
+    }
+  },
+
+  withToolCalls: {
+    text: 'I will check the weather for you.',
+    toolCalls: [
+      {
+        toolCallId: 'call_456',
+        toolName: 'getWeather',
+        args: { location: 'New York, NY', unit: 'celsius' }
+      }
+    ],
+    finishReason: 'tool-calls' as const,
+    usage: {
+      promptTokens: 25,
+      completionTokens: 12,
+      totalTokens: 37
+    }
+  },
+
+  withWarnings: {
+    text: 'Response with warnings.',
+    finishReason: 'stop' as const,
+    usage: {
+      promptTokens: 10,
+      completionTokens: 5,
+      totalTokens: 15
+    },
+    warnings: [
+      {
+        type: 'unsupported-setting' as const,
+        message: 'Temperature parameter not supported for this model'
+      }
+    ]
+  }
+}
+
+/**
+ * Mock image generation responses
+ */
+export const mockImageResponses = {
+  single: {
+    image: {
+      base64: 'iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mNk+M9QDwADhgGAWjR9awAAAABJRU5ErkJggg==',
+      uint8Array: new Uint8Array([137, 80, 78, 71, 13, 10, 26, 10, 0, 0, 0, 13, 73, 72, 68, 82]),
+      mimeType: 'image/png' as const
+    },
+    warnings: []
+  },
+
+  multiple: {
+    images: [
+      {
+        base64: 'iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mNk+M9QDwADhgGAWjR9awAAAABJRU5ErkJggg==',
+        uint8Array: new Uint8Array([137, 80, 78, 71]),
+        mimeType: 'image/png' as const
+      },
+      {
+        base64: 'iVBORw0KGgoAAAANSUhEUgAAAAIAAAACCAYAAABytg0kAAAAEklEQVR42mNk+M9QzwAEjDAGACCKAgdZ9zImAAAAAElFTkSuQmCC',
+        uint8Array: new Uint8Array([137, 80, 78, 71]),
+        mimeType: 'image/png' as const
+      }
+    ],
+    warnings: []
+  },
+
+  withProviderMetadata: {
+    image: {
+      base64: 'iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mNk+M9QDwADhgGAWjR9awAAAABJRU5ErkJggg==',
+      uint8Array: new Uint8Array([137, 80, 78, 71]),
+      mimeType: 'image/png' as const
+    },
+    providerMetadata: {
+      openai: {
+        images: [
+          {
+            revisedPrompt: 'A detailed and enhanced version of the original prompt'
+          }
+        ]
+      }
+    },
+    warnings: []
+  }
+}
+
+/**
+ * Mock error responses
+ */
+export const mockErrors = {
+  invalidApiKey: {
+    name: 'APIError',
+    message: 'Invalid API key provided',
+    statusCode: 401
+  },
+
+  rateLimitExceeded: {
+    name: 'RateLimitError',
+    message: 'Rate limit exceeded. Please try again later.',
+    statusCode: 429,
+    headers: {
+      'retry-after': '60'
+    }
+  },
+
+  modelNotFound: {
+    name: 'ModelNotFoundError',
+    message: 'The requested model was not found',
+    statusCode: 404
+  },
+
+  contextLengthExceeded: {
+    name: 'ContextLengthError',
+    message: "This model's maximum context length is 4096 tokens",
+    statusCode: 400
+  },
+
+  timeout: {
+    name: 'TimeoutError',
+    message: 'Request timed out after 30000ms',
+    code: 'ETIMEDOUT'
+  },
+
+  networkError: {
+    name: 'NetworkError',
+    message: 'Network connection failed',
+    code: 'ECONNREFUSED'
+  }
+}
@@ -0,0 +1,329 @@
+/**
+ * Provider-Specific Test Utilities
+ * Helper functions for testing individual providers with all their parameters
+ */
+
+import type { Tool } from 'ai'
+import { expect } from 'vitest'
+
+/**
+ * Provider parameter configurations for comprehensive testing
+ */
+export const providerParameterMatrix = {
+  openai: {
+    models: ['gpt-4', 'gpt-4-turbo', 'gpt-3.5-turbo', 'gpt-4o'],
+    parameters: {
+      temperature: [0, 0.5, 0.7, 1.0, 1.5, 2.0],
+      maxTokens: [100, 500, 1000, 2000, 4000],
+      topP: [0.1, 0.5, 0.9, 1.0],
+      frequencyPenalty: [-2.0, -1.0, 0, 1.0, 2.0],
+      presencePenalty: [-2.0, -1.0, 0, 1.0, 2.0],
+      stop: [undefined, ['stop'], ['STOP', 'END']],
+      seed: [undefined, 12345, 67890],
+      responseFormat: [undefined, { type: 'json_object' as const }],
+      user: [undefined, 'test-user-123']
+    },
+    toolChoice: ['auto', 'required', 'none', { type: 'function' as const, name: 'getWeather' }],
+    parallelToolCalls: [true, false]
+  },
+
+  anthropic: {
+    models: ['claude-3-5-sonnet-20241022', 'claude-3-opus-20240229', 'claude-3-haiku-20240307'],
+    parameters: {
+      temperature: [0, 0.5, 1.0],
+      maxTokens: [100, 1000, 4000, 8000],
+      topP: [0.1, 0.5, 0.9, 1.0],
+      topK: [undefined, 1, 5, 10, 40],
+      stop: [undefined, ['Human:', 'Assistant:']],
+      metadata: [undefined, { userId: 'test-123' }]
+    },
+    toolChoice: ['auto', 'any', { type: 'tool' as const, name: 'getWeather' }]
+  },
+
+  google: {
+    models: ['gemini-2.0-flash-exp', 'gemini-1.5-pro', 'gemini-1.5-flash'],
+    parameters: {
+      temperature: [0, 0.5, 0.9, 1.0],
+      maxTokens: [100, 1000, 2000, 8000],
+      topP: [0.1, 0.5, 0.95, 1.0],
+      topK: [undefined, 1, 16, 40],
+      stopSequences: [undefined, ['END'], ['STOP', 'TERMINATE']]
+    },
+    safetySettings: [
+      undefined,
+      [
+        { category: 'HARM_CATEGORY_HARASSMENT', threshold: 'BLOCK_MEDIUM_AND_ABOVE' },
+        { category: 'HARM_CATEGORY_HATE_SPEECH', threshold: 'BLOCK_ONLY_HIGH' }
+      ]
+    ]
+  },
+
+  xai: {
+    models: ['grok-2-latest', 'grok-2-1212'],
+    parameters: {
+      temperature: [0, 0.5, 1.0, 1.5],
+      maxTokens: [100, 500, 2000, 4000],
+      topP: [0.1, 0.5, 0.9, 1.0],
+      stop: [undefined, ['STOP'], ['END', 'TERMINATE']],
+      seed: [undefined, 12345]
+    }
+  },
+
+  deepseek: {
+    models: ['deepseek-chat', 'deepseek-coder'],
+    parameters: {
+      temperature: [0, 0.5, 1.0],
+      maxTokens: [100, 1000, 4000],
+      topP: [0.1, 0.5, 0.95],
+      frequencyPenalty: [0, 0.5, 1.0],
+      presencePenalty: [0, 0.5, 1.0],
+      stop: [undefined, ['```'], ['END']]
+    }
+  },
+
+  azure: {
+    deployments: ['gpt-4-deployment', 'gpt-35-turbo-deployment'],
+    parameters: {
+      temperature: [0, 0.7, 1.0],
+      maxTokens: [100, 1000, 2000],
+      topP: [0.1, 0.5, 0.95],
+      frequencyPenalty: [0, 1.0],
+      presencePenalty: [0, 1.0],
+      stop: [undefined, ['STOP']]
+    }
+  }
+} as const
+
+/**
+ * Creates test cases for all parameter combinations
+ */
+export function generateParameterTestCases<T extends Record<string, any[]>>(
+  params: T,
+  maxCombinations = 50
+): Array<Partial<{ [K in keyof T]: T[K][number] }>> {
+  const keys = Object.keys(params) as Array<keyof T>
+  const testCases: Array<Partial<{ [K in keyof T]: T[K][number] }>> = []
+
+  // Generate combinations using sampling strategy for large parameter spaces
+  const totalCombinations = keys.reduce((acc, key) => acc * params[key].length, 1)
+
+  if (totalCombinations <= maxCombinations) {
+    // Generate all combinations if total is small
+    generateAllCombinations(params, keys, 0, {}, testCases)
+  } else {
+    // Sample diverse combinations if total is large
+    generateSampledCombinations(params, keys, maxCombinations, testCases)
+  }
+
+  return testCases
+}
+
+function generateAllCombinations<T extends Record<string, any[]>>(
+  params: T,
+  keys: Array<keyof T>,
+  index: number,
+  current: Partial<{ [K in keyof T]: T[K][number] }>,
+  results: Array<Partial<{ [K in keyof T]: T[K][number] }>>
+) {
+  if (index === keys.length) {
+    results.push({ ...current })
+    return
+  }
+
+  const key = keys[index]
+  for (const value of params[key]) {
+    generateAllCombinations(params, keys, index + 1, { ...current, [key]: value }, results)
+  }
+}
+
+function generateSampledCombinations<T extends Record<string, any[]>>(
+  params: T,
+  keys: Array<keyof T>,
+  count: number,
+  results: Array<Partial<{ [K in keyof T]: T[K][number] }>>
+) {
+  // Generate edge cases first (min/max values)
+  const edgeCase1: any = {}
+  const edgeCase2: any = {}
+
+  for (const key of keys) {
+    edgeCase1[key] = params[key][0]
+    edgeCase2[key] = params[key][params[key].length - 1]
+  }
+
+  results.push(edgeCase1, edgeCase2)
+
+  // Generate random combinations for the rest
+  for (let i = results.length; i < count; i++) {
+    const combination: any = {}
+    for (const key of keys) {
+      const values = params[key]
+      combination[key] = values[Math.floor(Math.random() * values.length)]
+    }
+    results.push(combination)
+  }
+}
+
+/**
+ * Validates that all provider-specific parameters are correctly passed through
+ */
+export function validateProviderParams(providerId: string, actualParams: any, expectedParams: any): void {
+  const requiredFields: Record<string, string[]> = {
+    openai: ['model', 'messages'],
+    anthropic: ['model', 'messages'],
+    google: ['model', 'contents'],
+    xai: ['model', 'messages'],
+    deepseek: ['model', 'messages'],
+    azure: ['messages']
+  }
+
+  const fields = requiredFields[providerId] || ['model', 'messages']
+
+  for (const field of fields) {
+    expect(actualParams).toHaveProperty(field)
+  }
+
+  // Validate optional parameters if they were provided
+  const optionalParams = ['temperature', 'max_tokens', 'top_p', 'stop', 'tools']
+
+  for (const param of optionalParams) {
+    if (expectedParams[param] !== undefined) {
+      expect(actualParams[param]).toEqual(expectedParams[param])
+    }
+  }
+}
+
+/**
+ * Creates a comprehensive test suite for a provider
+ */
+// oxlint-disable-next-line no-unused-vars
+export function createProviderTestSuite(_providerId: string) {
+  return {
+    testBasicCompletion: async (executor: any, model: string) => {
+      const result = await executor.generateText({
+        model,
+        messages: [{ role: 'user' as const, content: 'Hello' }]
+      })
+
+      expect(result).toBeDefined()
+      expect(result.text).toBeDefined()
+      expect(typeof result.text).toBe('string')
+    },
+
+    testStreaming: async (executor: any, model: string) => {
+      const chunks: any[] = []
+      const result = await executor.streamText({
+        model,
+        messages: [{ role: 'user' as const, content: 'Hello' }]
+      })
+
+      for await (const chunk of result.textStream) {
+        chunks.push(chunk)
+      }
+
+      expect(chunks.length).toBeGreaterThan(0)
+    },
+
+    testTemperature: async (executor: any, model: string, temperatures: number[]) => {
+      for (const temperature of temperatures) {
+        const result = await executor.generateText({
+          model,
+          messages: [{ role: 'user' as const, content: 'Hello' }],
+          temperature
+        })
+
+        expect(result).toBeDefined()
+      }
+    },
+
+    testMaxTokens: async (executor: any, model: string, maxTokensValues: number[]) => {
+      for (const maxTokens of maxTokensValues) {
+        const result = await executor.generateText({
+          model,
+          messages: [{ role: 'user' as const, content: 'Hello' }],
+          maxTokens
+        })
+
+        expect(result).toBeDefined()
+        if (result.usage?.completionTokens) {
+          expect(result.usage.completionTokens).toBeLessThanOrEqual(maxTokens)
+        }
+      }
+    },
+
+    testToolCalling: async (executor: any, model: string, tools: Record<string, Tool>) => {
+      const result = await executor.generateText({
+        model,
+        messages: [{ role: 'user' as const, content: 'What is the weather in SF?' }],
+        tools
+      })
+
+      expect(result).toBeDefined()
+    },
+
+    testStopSequences: async (executor: any, model: string, stopSequences: string[][]) => {
+      for (const stop of stopSequences) {
+        const result = await executor.generateText({
+          model,
+          messages: [{ role: 'user' as const, content: 'Count to 10' }],
+          stop
+        })
+
+        expect(result).toBeDefined()
+      }
+    }
+  }
+}
+
+/**
+ * Generates test data for vision/multimodal testing
+ */
+export function createVisionTestData() {
+  return {
+    imageUrl: 'https://example.com/test-image.jpg',
+    base64Image:
+      'data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mNk+M9QDwADhgGAWjR9awAAAABJRU5ErkJggg==',
+    messages: [
+      {
+        role: 'user' as const,
+        content: [
+          { type: 'text' as const, text: 'What is in this image?' },
+          {
+            type: 'image' as const,
+            image:
+              'data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mNk+M9QDwADhgGAWjR9awAAAABJRU5ErkJggg=='
+          }
+        ]
+      }
+    ]
+  }
+}
+
+/**
+ * Creates mock responses for different finish reasons
+ */
+export function createFinishReasonMocks() {
+  return {
+    stop: {
+      text: 'Complete response.',
+      finishReason: 'stop' as const,
+      usage: { promptTokens: 10, completionTokens: 5, totalTokens: 15 }
+    },
+    length: {
+      text: 'Incomplete response due to',
+      finishReason: 'length' as const,
+      usage: { promptTokens: 10, completionTokens: 100, totalTokens: 110 }
+    },
+    'tool-calls': {
+      text: 'Calling tools',
+      finishReason: 'tool-calls' as const,
+      toolCalls: [{ toolCallId: 'call_1', toolName: 'getWeather', args: { location: 'SF' } }],
+      usage: { promptTokens: 10, completionTokens: 8, totalTokens: 18 }
+    },
+    'content-filter': {
+      text: '',
+      finishReason: 'content-filter' as const,
+      usage: { promptTokens: 10, completionTokens: 0, totalTokens: 10 }
+    }
+  }
+}
@@ -0,0 +1,291 @@
+/**
+ * Test Utilities
+ * Helper functions for testing AI Core functionality
+ */
+
+import { expect, vi } from 'vitest'
+
+import type { ProviderId } from '../fixtures/mock-providers'
+import { createMockImageModel, createMockLanguageModel, mockProviderConfigs } from '../fixtures/mock-providers'
+
+/**
+ * Creates a test provider with streaming support
+ */
+export function createTestStreamingProvider(chunks: any[]) {
+  return createMockLanguageModel({
+    doStream: vi.fn().mockReturnValue({
+      stream: (async function* () {
+        for (const chunk of chunks) {
+          yield chunk
+        }
+      })(),
+      rawCall: { rawPrompt: null, rawSettings: {} },
+      rawResponse: { headers: {} },
+      warnings: []
+    })
+  })
+}
+
+/**
+ * Creates a test provider that throws errors
+ */
+export function createErrorProvider(error: Error) {
+  return createMockLanguageModel({
+    doGenerate: vi.fn().mockRejectedValue(error),
+    doStream: vi.fn().mockImplementation(() => {
+      throw error
+    })
+  })
+}
+
+/**
+ * Collects all chunks from a stream
+ */
+export async function collectStreamChunks<T>(stream: AsyncIterable<T>): Promise<T[]> {
+  const chunks: T[] = []
+  for await (const chunk of stream) {
+    chunks.push(chunk)
+  }
+  return chunks
+}
+
+/**
+ * Waits for a specific number of milliseconds
+ */
+export function wait(ms: number): Promise<void> {
+  return new Promise((resolve) => setTimeout(resolve, ms))
+}
+
+/**
+ * Creates a mock abort controller that aborts after a delay
+ */
+export function createDelayedAbortController(delayMs: number): AbortController {
+  const controller = new AbortController()
+  setTimeout(() => controller.abort(), delayMs)
+  return controller
+}
+
+/**
+ * Asserts that a function throws an error with a specific message
+ */
+export async function expectError(fn: () => Promise<any>, expectedMessage?: string | RegExp): Promise<Error> {
+  try {
+    await fn()
+    throw new Error('Expected function to throw an error, but it did not')
+  } catch (error) {
+    if (expectedMessage) {
+      const message = (error as Error).message
+      if (typeof expectedMessage === 'string') {
+        if (!message.includes(expectedMessage)) {
+          throw new Error(`Expected error message to include "${expectedMessage}", but got "${message}"`)
+        }
+      } else {
+        if (!expectedMessage.test(message)) {
+          throw new Error(`Expected error message to match ${expectedMessage}, but got "${message}"`)
+        }
+      }
+    }
+    return error as Error
+  }
+}
+
+/**
+ * Creates a spy function that tracks calls and arguments
+ */
+export function createSpy<T extends (...args: any[]) => any>() {
+  const calls: Array<{ args: Parameters<T>; result?: ReturnType<T>; error?: Error }> = []
+
+  const spy = vi.fn((...args: Parameters<T>) => {
+    try {
+      const result = undefined as ReturnType<T>
+      calls.push({ args, result })
+      return result
+    } catch (error) {
+      calls.push({ args, error: error as Error })
+      throw error
+    }
+  })
+
+  return {
+    fn: spy,
+    calls,
+    getCalls: () => calls,
+    getCallCount: () => calls.length,
+    getLastCall: () => calls[calls.length - 1],
+    reset: () => {
+      calls.length = 0
+      spy.mockClear()
+    }
+  }
+}
+
+/**
+ * Validates provider configuration
+ */
+export function validateProviderConfig(providerId: ProviderId) {
+  const config = mockProviderConfigs[providerId]
+  if (!config) {
+    throw new Error(`No mock configuration found for provider: ${providerId}`)
+  }
+
+  if (!config.apiKey) {
+    throw new Error(`Provider ${providerId} is missing apiKey in mock config`)
+  }
+
+  return config
+}
+
+/**
+ * Creates a test context with common setup
+ */
+export function createTestContext() {
+  const mocks = {
+    languageModel: createMockLanguageModel(),
+    imageModel: createMockImageModel(),
+    providers: new Map<string, any>()
+  }
+
+  const cleanup = () => {
+    mocks.providers.clear()
+    vi.clearAllMocks()
+  }
+
+  return {
+    mocks,
+    cleanup
+  }
+}
+
+/**
+ * Measures execution time of an async function
+ */
+export async function measureTime<T>(fn: () => Promise<T>): Promise<{ result: T; duration: number }> {
+  const start = Date.now()
+  const result = await fn()
+  const duration = Date.now() - start
+  return { result, duration }
+}
+
+/**
+ * Retries a function until it succeeds or max attempts reached
+ */
+export async function retryUntilSuccess<T>(fn: () => Promise<T>, maxAttempts = 3, delayMs = 100): Promise<T> {
+  let lastError: Error | undefined
+
+  for (let attempt = 1; attempt <= maxAttempts; attempt++) {
+    try {
+      return await fn()
+    } catch (error) {
+      lastError = error as Error
+      if (attempt < maxAttempts) {
+        await wait(delayMs)
+      }
+    }
+  }
+
+  throw lastError || new Error('All retry attempts failed')
+}
+
+/**
+ * Creates a mock streaming response that emits chunks at intervals
+ */
+export function createTimedStream<T>(chunks: T[], intervalMs = 10) {
+  return {
+    async *[Symbol.asyncIterator]() {
+      for (const chunk of chunks) {
+        await wait(intervalMs)
+        yield chunk
+      }
+    }
+  }
+}
+
+/**
+ * Asserts that two objects are deeply equal, ignoring specified keys
+ */
+export function assertDeepEqualIgnoring<T extends Record<string, any>>(
+  actual: T,
+  expected: T,
+  ignoreKeys: string[] = []
+): void {
+  const filterKeys = (obj: T): Partial<T> => {
+    const filtered = { ...obj }
+    for (const key of ignoreKeys) {
+      delete filtered[key]
+    }
+    return filtered
+  }
+
+  const filteredActual = filterKeys(actual)
+  const filteredExpected = filterKeys(expected)
+
+  expect(filteredActual).toEqual(filteredExpected)
+}
+
+/**
+ * Creates a provider mock that simulates rate limiting
+ */
+export function createRateLimitedProvider(limitPerSecond: number) {
+  const calls: number[] = []
+
+  return createMockLanguageModel({
+    doGenerate: vi.fn().mockImplementation(async () => {
+      const now = Date.now()
+      calls.push(now)
+
+      // Remove calls older than 1 second
+      const recentCalls = calls.filter((time) => now - time < 1000)
+
+      if (recentCalls.length > limitPerSecond) {
+        throw new Error('Rate limit exceeded')
+      }
+
+      return {
+        text: 'Rate limited response',
+        finishReason: 'stop' as const,
+        usage: { promptTokens: 10, completionTokens: 5, totalTokens: 15 },
+        rawCall: { rawPrompt: null, rawSettings: {} },
+        rawResponse: { headers: {} },
+        warnings: []
+      }
+    })
+  })
+}
+
+/**
+ * Validates streaming response structure
+ */
+export function validateStreamChunk(chunk: any): void {
+  expect(chunk).toBeDefined()
+  expect(chunk).toHaveProperty('type')
+
+  if (chunk.type === 'text-delta') {
+    expect(chunk).toHaveProperty('textDelta')
+    expect(typeof chunk.textDelta).toBe('string')
+  } else if (chunk.type === 'finish') {
+    expect(chunk).toHaveProperty('finishReason')
+    expect(chunk).toHaveProperty('usage')
+  } else if (chunk.type === 'tool-call') {
+    expect(chunk).toHaveProperty('toolCallId')
+    expect(chunk).toHaveProperty('toolName')
+    expect(chunk).toHaveProperty('args')
+  }
+}
+
+/**
+ * Creates a test logger that captures log messages
+ */
+export function createTestLogger() {
+  const logs: Array<{ level: string; message: string; meta?: any }> = []
+
+  return {
+    info: (message: string, meta?: any) => logs.push({ level: 'info', message, meta }),
+    warn: (message: string, meta?: any) => logs.push({ level: 'warn', message, meta }),
+    error: (message: string, meta?: any) => logs.push({ level: 'error', message, meta }),
+    debug: (message: string, meta?: any) => logs.push({ level: 'debug', message, meta }),
+    getLogs: () => logs,
+    clear: () => {
+      logs.length = 0
+    }
+  }
+}
@@ -0,0 +1,12 @@
+/**
+ * Test Infrastructure Exports
+ * Central export point for all test utilities, fixtures, and helpers
+ */
+
+// Fixtures
+export * from './fixtures/mock-providers'
+export * from './fixtures/mock-responses'
+
+// Helpers
+export * from './helpers/provider-test-utils'
+export * from './helpers/test-utils'
@@ -4,12 +4,7 @@
 */
 export const BUILT_IN_PLUGIN_PREFIX = 'built-in:'

-export { googleToolsPlugin } from './googleToolsPlugin'
-export { createLoggingPlugin } from './logging'
-export { createPromptToolUsePlugin } from './toolUsePlugin/promptToolUsePlugin'
-export type {
-  PromptToolUseConfig,
-  ToolUseRequestContext,
-  ToolUseResult
-} from './toolUsePlugin/type'
-export { webSearchPlugin, type WebSearchPluginConfig } from './webSearchPlugin'
+export * from './googleToolsPlugin'
+export * from './toolUsePlugin/promptToolUsePlugin'
+export * from './toolUsePlugin/type'
+export * from './webSearchPlugin'
@@ -32,7 +32,7 @@ export const webSearchPlugin = (config: WebSearchPluginConfig = DEFAULT_WEB_SEAR
  })

 // 导出类型定义供开发者使用
-export type { WebSearchPluginConfig, WebSearchToolOutputSchema } from './helper'
+export * from './helper'

 // 默认导出
 export default webSearchPlugin
@@ -44,7 +44,7 @@ export {
 // ==================== 基础数据和类型 ====================

 // 基础Provider数据源
-export { baseProviderIds, baseProviders } from './schemas'
+export { baseProviderIds, baseProviders, isBaseProvider } from './schemas'

 // 类型定义和Schema
 export type {
@@ -7,7 +7,6 @@ import { createAzure } from '@ai-sdk/azure'
 import { type AzureOpenAIProviderSettings } from '@ai-sdk/azure'
 import { createDeepSeek } from '@ai-sdk/deepseek'
 import { createGoogleGenerativeAI } from '@ai-sdk/google'
-import { createHuggingFace } from '@ai-sdk/huggingface'
 import { createOpenAI, type OpenAIProviderSettings } from '@ai-sdk/openai'
 import { createOpenAICompatible } from '@ai-sdk/openai-compatible'
 import type { LanguageModelV2 } from '@ai-sdk/provider'
@@ -33,8 +32,7 @@ export const baseProviderIds = [
  'deepseek',
  'openrouter',
  'cherryin',
-  'cherryin-chat',
-  'huggingface'
+  'cherryin-chat'
 ] as const

 /**
@@ -158,12 +156,6 @@ export const baseProviders = [
      })
    },
    supportsImageGeneration: true
-  },
-  {
-    id: 'huggingface',
-    name: 'HuggingFace',
-    creator: createHuggingFace,
-    supportsImageGeneration: true
  }
 ] as const satisfies BaseProvider[]

@@ -0,0 +1,499 @@
+/**
+ * RuntimeExecutor.generateText Comprehensive Tests
+ * Tests non-streaming text generation across all providers with various parameters
+ */
+
+import { generateText } from 'ai'
+import { beforeEach, describe, expect, it, vi } from 'vitest'
+
+import {
+  createMockLanguageModel,
+  mockCompleteResponses,
+  mockProviderConfigs,
+  testMessages,
+  testTools
+} from '../../../__tests__'
+import type { AiPlugin } from '../../plugins'
+import { globalRegistryManagement } from '../../providers/RegistryManagement'
+import { RuntimeExecutor } from '../executor'
+
+// Mock AI SDK
+vi.mock('ai', () => ({
+  generateText: vi.fn()
+}))
+
+vi.mock('../../providers/RegistryManagement', () => ({
+  globalRegistryManagement: {
+    languageModel: vi.fn()
+  },
+  DEFAULT_SEPARATOR: '|'
+}))
+
+describe('RuntimeExecutor.generateText', () => {
+  let executor: RuntimeExecutor<'openai'>
+  let mockLanguageModel: any
+
+  beforeEach(() => {
+    vi.clearAllMocks()
+
+    executor = RuntimeExecutor.create('openai', mockProviderConfigs.openai)
+
+    mockLanguageModel = createMockLanguageModel({
+      provider: 'openai',
+      modelId: 'gpt-4'
+    })
+
+    vi.mocked(globalRegistryManagement.languageModel).mockReturnValue(mockLanguageModel)
+    vi.mocked(generateText).mockResolvedValue(mockCompleteResponses.simple as any)
+  })
+
+  describe('Basic Functionality', () => {
+    it('should generate text with minimal parameters', async () => {
+      const result = await executor.generateText({
+        model: 'gpt-4',
+        messages: testMessages.simple
+      })
+
+      expect(generateText).toHaveBeenCalledWith({
+        model: mockLanguageModel,
+        messages: testMessages.simple
+      })
+
+      expect(result.text).toBe('This is a simple response.')
+      expect(result.finishReason).toBe('stop')
+      expect(result.usage).toBeDefined()
+    })
+
+    it('should generate with system messages', async () => {
+      await executor.generateText({
+        model: 'gpt-4',
+        messages: testMessages.withSystem
+      })
+
+      expect(generateText).toHaveBeenCalledWith({
+        model: mockLanguageModel,
+        messages: testMessages.withSystem
+      })
+    })
+
+    it('should generate with conversation history', async () => {
+      await executor.generateText({
+        model: 'gpt-4',
+        messages: testMessages.conversation
+      })
+
+      expect(generateText).toHaveBeenCalledWith(
+        expect.objectContaining({
+          messages: testMessages.conversation
+        })
+      )
+    })
+  })
+
+  describe('All Parameter Combinations', () => {
+    it('should support all parameters together', async () => {
+      await executor.generateText({
+        model: 'gpt-4',
+        messages: testMessages.simple,
+        temperature: 0.7,
+        maxOutputTokens: 500,
+        topP: 0.9,
+        frequencyPenalty: 0.5,
+        presencePenalty: 0.3,
+        stopSequences: ['STOP'],
+        seed: 12345
+      })
+
+      expect(generateText).toHaveBeenCalledWith(
+        expect.objectContaining({
+          temperature: 0.7,
+          maxOutputTokens: 500,
+          topP: 0.9,
+          frequencyPenalty: 0.5,
+          presencePenalty: 0.3,
+          stopSequences: ['STOP'],
+          seed: 12345
+        })
+      )
+    })
+
+    it('should support partial parameters', async () => {
+      await executor.generateText({
+        model: 'gpt-4',
+        messages: testMessages.simple,
+        temperature: 0.5,
+        maxOutputTokens: 100
+      })
+
+      expect(generateText).toHaveBeenCalledWith(
+        expect.objectContaining({
+          temperature: 0.5,
+          maxOutputTokens: 100
+        })
+      )
+    })
+  })
+
+  describe('Tool Calling', () => {
+    beforeEach(() => {
+      vi.mocked(generateText).mockResolvedValue(mockCompleteResponses.withToolCalls as any)
+    })
+
+    it('should support tool calling', async () => {
+      const result = await executor.generateText({
+        model: 'gpt-4',
+        messages: testMessages.toolUse,
+        tools: testTools
+      })
+
+      expect(generateText).toHaveBeenCalledWith(
+        expect.objectContaining({
+          tools: testTools
+        })
+      )
+
+      expect(result.toolCalls).toBeDefined()
+      expect(result.toolCalls).toHaveLength(1)
+    })
+
+    it('should support toolChoice auto', async () => {
+      await executor.generateText({
+        model: 'gpt-4',
+        messages: testMessages.toolUse,
+        tools: testTools,
+        toolChoice: 'auto'
+      })
+
+      expect(generateText).toHaveBeenCalledWith(
+        expect.objectContaining({
+          toolChoice: 'auto'
+        })
+      )
+    })
+
+    it('should support toolChoice required', async () => {
+      await executor.generateText({
+        model: 'gpt-4',
+        messages: testMessages.toolUse,
+        tools: testTools,
+        toolChoice: 'required'
+      })
+
+      expect(generateText).toHaveBeenCalledWith(
+        expect.objectContaining({
+          toolChoice: 'required'
+        })
+      )
+    })
+
+    it('should support toolChoice none', async () => {
+      vi.mocked(generateText).mockResolvedValue(mockCompleteResponses.simple as any)
+
+      await executor.generateText({
+        model: 'gpt-4',
+        messages: testMessages.simple,
+        tools: testTools,
+        toolChoice: 'none'
+      })
+
+      expect(generateText).toHaveBeenCalledWith(
+        expect.objectContaining({
+          toolChoice: 'none'
+        })
+      )
+    })
+
+    it('should support specific tool selection', async () => {
+      await executor.generateText({
+        model: 'gpt-4',
+        messages: testMessages.toolUse,
+        tools: testTools,
+        toolChoice: {
+          type: 'tool',
+          toolName: 'getWeather'
+        }
+      })
+
+      expect(generateText).toHaveBeenCalledWith(
+        expect.objectContaining({
+          toolChoice: {
+            type: 'tool',
+            toolName: 'getWeather'
+          }
+        })
+      )
+    })
+  })
+
+  describe('Multiple Providers', () => {
+    it('should work with Anthropic provider', async () => {
+      const anthropicExecutor = RuntimeExecutor.create('anthropic', mockProviderConfigs.anthropic)
+
+      const anthropicModel = createMockLanguageModel({
+        provider: 'anthropic',
+        modelId: 'claude-3-5-sonnet-20241022'
+      })
+
+      vi.mocked(globalRegistryManagement.languageModel).mockReturnValue(anthropicModel)
+
+      await anthropicExecutor.generateText({
+        model: 'claude-3-5-sonnet-20241022',
+        messages: testMessages.simple
+      })
+
+      expect(globalRegistryManagement.languageModel).toHaveBeenCalledWith('anthropic|claude-3-5-sonnet-20241022')
+    })
+
+    it('should work with Google provider', async () => {
+      const googleExecutor = RuntimeExecutor.create('google', mockProviderConfigs.google)
+
+      const googleModel = createMockLanguageModel({
+        provider: 'google',
+        modelId: 'gemini-2.0-flash-exp'
+      })
+
+      vi.mocked(globalRegistryManagement.languageModel).mockReturnValue(googleModel)
+
+      await googleExecutor.generateText({
+        model: 'gemini-2.0-flash-exp',
+        messages: testMessages.simple
+      })
+
+      expect(globalRegistryManagement.languageModel).toHaveBeenCalledWith('google|gemini-2.0-flash-exp')
+    })
+
+    it('should work with xAI provider', async () => {
+      const xaiExecutor = RuntimeExecutor.create('xai', mockProviderConfigs.xai)
+
+      const xaiModel = createMockLanguageModel({
+        provider: 'xai',
+        modelId: 'grok-2-latest'
+      })
+
+      vi.mocked(globalRegistryManagement.languageModel).mockReturnValue(xaiModel)
+
+      await xaiExecutor.generateText({
+        model: 'grok-2-latest',
+        messages: testMessages.simple
+      })
+
+      expect(globalRegistryManagement.languageModel).toHaveBeenCalledWith('xai|grok-2-latest')
+    })
+
+    it('should work with DeepSeek provider', async () => {
+      const deepseekExecutor = RuntimeExecutor.create('deepseek', mockProviderConfigs.deepseek)
+
+      const deepseekModel = createMockLanguageModel({
+        provider: 'deepseek',
+        modelId: 'deepseek-chat'
+      })
+
+      vi.mocked(globalRegistryManagement.languageModel).mockReturnValue(deepseekModel)
+
+      await deepseekExecutor.generateText({
+        model: 'deepseek-chat',
+        messages: testMessages.simple
+      })
+
+      expect(globalRegistryManagement.languageModel).toHaveBeenCalledWith('deepseek|deepseek-chat')
+    })
+  })
+
+  describe('Plugin Integration', () => {
+    it('should execute all plugin hooks', async () => {
+      const pluginCalls: string[] = []
+
+      const testPlugin: AiPlugin = {
+        name: 'test-plugin',
+        onRequestStart: vi.fn(async () => {
+          pluginCalls.push('onRequestStart')
+        }),
+        transformParams: vi.fn(async (params) => {
+          pluginCalls.push('transformParams')
+          return { ...params, temperature: 0.8 }
+        }),
+        transformResult: vi.fn(async (result) => {
+          pluginCalls.push('transformResult')
+          return { ...result, text: result.text + ' [modified]' }
+        }),
+        onRequestEnd: vi.fn(async () => {
+          pluginCalls.push('onRequestEnd')
+        })
+      }
+
+      const executorWithPlugin = RuntimeExecutor.create('openai', mockProviderConfigs.openai, [testPlugin])
+
+      const result = await executorWithPlugin.generateText({
+        model: 'gpt-4',
+        messages: testMessages.simple
+      })
+
+      expect(pluginCalls).toEqual(['onRequestStart', 'transformParams', 'transformResult', 'onRequestEnd'])
+
+      // Verify transformed parameters
+      expect(generateText).toHaveBeenCalledWith(
+        expect.objectContaining({
+          temperature: 0.8
+        })
+      )
+
+      // Verify transformed result
+      expect(result.text).toContain('[modified]')
+    })
+
+    it('should handle multiple plugins in order', async () => {
+      const pluginOrder: string[] = []
+
+      const plugin1: AiPlugin = {
+        name: 'plugin-1',
+        transformParams: vi.fn(async (params) => {
+          pluginOrder.push('plugin-1')
+          return { ...params, temperature: 0.5 }
+        })
+      }
+
+      const plugin2: AiPlugin = {
+        name: 'plugin-2',
+        transformParams: vi.fn(async (params) => {
+          pluginOrder.push('plugin-2')
+          return { ...params, maxTokens: 200 }
+        })
+      }
+
+      const executorWithPlugins = RuntimeExecutor.create('openai', mockProviderConfigs.openai, [plugin1, plugin2])
+
+      await executorWithPlugins.generateText({
+        model: 'gpt-4',
+        messages: testMessages.simple
+      })
+
+      expect(pluginOrder).toEqual(['plugin-1', 'plugin-2'])
+
+      expect(generateText).toHaveBeenCalledWith(
+        expect.objectContaining({
+          temperature: 0.5,
+          maxTokens: 200
+        })
+      )
+    })
+  })
+
+  describe('Error Handling', () => {
+    it('should handle API errors', async () => {
+      const error = new Error('API request failed')
+      vi.mocked(generateText).mockRejectedValue(error)
+
+      await expect(
+        executor.generateText({
+          model: 'gpt-4',
+          messages: testMessages.simple
+        })
+      ).rejects.toThrow('API request failed')
+    })
+
+    it('should execute onError plugin hook', async () => {
+      const error = new Error('Generation failed')
+      vi.mocked(generateText).mockRejectedValue(error)
+
+      const errorPlugin: AiPlugin = {
+        name: 'error-handler',
+        onError: vi.fn()
+      }
+
+      const executorWithPlugin = RuntimeExecutor.create('openai', mockProviderConfigs.openai, [errorPlugin])
+
+      await expect(
+        executorWithPlugin.generateText({
+          model: 'gpt-4',
+          messages: testMessages.simple
+        })
+      ).rejects.toThrow('Generation failed')
+
+      expect(errorPlugin.onError).toHaveBeenCalledWith(
+        error,
+        expect.objectContaining({
+          providerId: 'openai',
+          modelId: 'gpt-4'
+        })
+      )
+    })
+
+    it('should handle model not found error', async () => {
+      const error = new Error('Model not found: invalid-model')
+      vi.mocked(globalRegistryManagement.languageModel).mockImplementation(() => {
+        throw error
+      })
+
+      await expect(
+        executor.generateText({
+          model: 'invalid-model',
+          messages: testMessages.simple
+        })
+      ).rejects.toThrow('Model not found')
+    })
+  })
+
+  describe('Usage and Metadata', () => {
+    it('should return usage information', async () => {
+      const result = await executor.generateText({
+        model: 'gpt-4',
+        messages: testMessages.simple
+      })
+
+      expect(result.usage).toBeDefined()
+      expect(result.usage.inputTokens).toBe(15)
+      expect(result.usage.outputTokens).toBe(8)
+      expect(result.usage.totalTokens).toBe(23)
+    })
+
+    it('should handle warnings', async () => {
+      vi.mocked(generateText).mockResolvedValue(mockCompleteResponses.withWarnings as any)
+
+      const result = await executor.generateText({
+        model: 'gpt-4',
+        messages: testMessages.simple,
+        temperature: 2.5 // Unsupported value
+      })
+
+      expect(result.warnings).toBeDefined()
+      expect(result.warnings).toHaveLength(1)
+      expect(result.warnings![0].type).toBe('unsupported-setting')
+    })
+  })
+
+  describe('Abort Signal', () => {
+    it('should support abort signal', async () => {
+      const abortController = new AbortController()
+
+      await executor.generateText({
+        model: 'gpt-4',
+        messages: testMessages.simple,
+        abortSignal: abortController.signal
+      })
+
+      expect(generateText).toHaveBeenCalledWith(
+        expect.objectContaining({
+          abortSignal: abortController.signal
+        })
+      )
+    })
+
+    it('should handle aborted request', async () => {
+      const abortError = new Error('Request aborted')
+      abortError.name = 'AbortError'
+
+      vi.mocked(generateText).mockRejectedValue(abortError)
+
+      const abortController = new AbortController()
+      abortController.abort()
+
+      await expect(
+        executor.generateText({
+          model: 'gpt-4',
+          messages: testMessages.simple,
+          abortSignal: abortController.signal
+        })
+      ).rejects.toThrow('Request aborted')
+    })
+  })
+})
@@ -0,0 +1,525 @@
+/**
+ * RuntimeExecutor.streamText Comprehensive Tests
+ * Tests streaming text generation across all providers with various parameters
+ */
+
+import { streamText } from 'ai'
+import { beforeEach, describe, expect, it, vi } from 'vitest'
+
+import { collectStreamChunks, createMockLanguageModel, mockProviderConfigs, testMessages } from '../../../__tests__'
+import type { AiPlugin } from '../../plugins'
+import { globalRegistryManagement } from '../../providers/RegistryManagement'
+import { RuntimeExecutor } from '../executor'
+
+// Mock AI SDK
+vi.mock('ai', () => ({
+  streamText: vi.fn()
+}))
+
+vi.mock('../../providers/RegistryManagement', () => ({
+  globalRegistryManagement: {
+    languageModel: vi.fn()
+  },
+  DEFAULT_SEPARATOR: '|'
+}))
+
+describe('RuntimeExecutor.streamText', () => {
+  let executor: RuntimeExecutor<'openai'>
+  let mockLanguageModel: any
+
+  beforeEach(() => {
+    vi.clearAllMocks()
+
+    executor = RuntimeExecutor.create('openai', mockProviderConfigs.openai)
+
+    mockLanguageModel = createMockLanguageModel({
+      provider: 'openai',
+      modelId: 'gpt-4'
+    })
+
+    vi.mocked(globalRegistryManagement.languageModel).mockReturnValue(mockLanguageModel)
+  })
+
+  describe('Basic Functionality', () => {
+    it('should stream text with minimal parameters', async () => {
+      const mockStream = {
+        textStream: (async function* () {
+          yield 'Hello'
+          yield ' '
+          yield 'World'
+        })(),
+        fullStream: (async function* () {
+          yield { type: 'text-delta', textDelta: 'Hello' }
+          yield { type: 'text-delta', textDelta: ' ' }
+          yield { type: 'text-delta', textDelta: 'World' }
+        })(),
+        usage: Promise.resolve({ promptTokens: 5, completionTokens: 3, totalTokens: 8 })
+      }
+
+      vi.mocked(streamText).mockResolvedValue(mockStream as any)
+
+      const result = await executor.streamText({
+        model: 'gpt-4',
+        messages: testMessages.simple
+      })
+
+      expect(streamText).toHaveBeenCalledWith({
+        model: mockLanguageModel,
+        messages: testMessages.simple
+      })
+
+      const chunks = await collectStreamChunks(result.textStream)
+      expect(chunks).toEqual(['Hello', ' ', 'World'])
+    })
+
+    it('should stream with system messages', async () => {
+      const mockStream = {
+        textStream: (async function* () {
+          yield 'Response'
+        })(),
+        fullStream: (async function* () {
+          yield { type: 'text-delta', textDelta: 'Response' }
+        })()
+      }
+
+      vi.mocked(streamText).mockResolvedValue(mockStream as any)
+
+      await executor.streamText({
+        model: 'gpt-4',
+        messages: testMessages.withSystem
+      })
+
+      expect(streamText).toHaveBeenCalledWith({
+        model: mockLanguageModel,
+        messages: testMessages.withSystem
+      })
+    })
+
+    it('should stream multi-turn conversations', async () => {
+      const mockStream = {
+        textStream: (async function* () {
+          yield 'Multi-turn response'
+        })(),
+        fullStream: (async function* () {
+          yield { type: 'text-delta', textDelta: 'Multi-turn response' }
+        })()
+      }
+
+      vi.mocked(streamText).mockResolvedValue(mockStream as any)
+
+      await executor.streamText({
+        model: 'gpt-4',
+        messages: testMessages.multiTurn
+      })
+
+      expect(streamText).toHaveBeenCalled()
+      expect(streamText).toHaveBeenCalledWith(
+        expect.objectContaining({
+          messages: testMessages.multiTurn
+        })
+      )
+    })
+  })
+
+  describe('Temperature Parameter', () => {
+    const temperatures = [0, 0.3, 0.5, 0.7, 0.9, 1.0, 1.5, 2.0]
+
+    it.each(temperatures)('should support temperature=%s', async (temperature) => {
+      const mockStream = {
+        textStream: (async function* () {
+          yield 'Response'
+        })(),
+        fullStream: (async function* () {
+          yield { type: 'text-delta', textDelta: 'Response' }
+        })()
+      }
+
+      vi.mocked(streamText).mockResolvedValue(mockStream as any)
+
+      await executor.streamText({
+        model: 'gpt-4',
+        messages: testMessages.simple,
+        temperature
+      })
+
+      expect(streamText).toHaveBeenCalledWith(
+        expect.objectContaining({
+          temperature
+        })
+      )
+    })
+  })
+
+  describe('Max Tokens Parameter', () => {
+    const maxTokensValues = [10, 50, 100, 500, 1000, 2000, 4000]
+
+    it.each(maxTokensValues)('should support maxTokens=%s', async (maxTokens) => {
+      const mockStream = {
+        textStream: (async function* () {
+          yield 'Response'
+        })(),
+        fullStream: (async function* () {
+          yield { type: 'text-delta', textDelta: 'Response' }
+        })()
+      }
+
+      vi.mocked(streamText).mockResolvedValue(mockStream as any)
+
+      await executor.streamText({
+        model: 'gpt-4',
+        messages: testMessages.simple,
+        maxOutputTokens: maxTokens
+      })
+
+      expect(streamText).toHaveBeenCalledWith(
+        expect.objectContaining({
+          maxTokens
+        })
+      )
+    })
+  })
+
+  describe('Top P Parameter', () => {
+    const topPValues = [0.1, 0.3, 0.5, 0.7, 0.9, 0.95, 1.0]
+
+    it.each(topPValues)('should support topP=%s', async (topP) => {
+      const mockStream = {
+        textStream: (async function* () {
+          yield 'Response'
+        })(),
+        fullStream: (async function* () {
+          yield { type: 'text-delta', textDelta: 'Response' }
+        })()
+      }
+
+      vi.mocked(streamText).mockResolvedValue(mockStream as any)
+
+      await executor.streamText({
+        model: 'gpt-4',
+        messages: testMessages.simple,
+        topP
+      })
+
+      expect(streamText).toHaveBeenCalledWith(
+        expect.objectContaining({
+          topP
+        })
+      )
+    })
+  })
+
+  describe('Frequency and Presence Penalty', () => {
+    it('should support frequency penalty', async () => {
+      const penalties = [-2.0, -1.0, 0, 0.5, 1.0, 1.5, 2.0]
+
+      for (const frequencyPenalty of penalties) {
+        vi.clearAllMocks()
+
+        const mockStream = {
+          textStream: (async function* () {
+            yield 'Response'
+          })(),
+          fullStream: (async function* () {
+            yield { type: 'text-delta', textDelta: 'Response' }
+          })()
+        }
+
+        vi.mocked(streamText).mockResolvedValue(mockStream as any)
+
+        await executor.streamText({
+          model: 'gpt-4',
+          messages: testMessages.simple,
+          frequencyPenalty
+        })
+
+        expect(streamText).toHaveBeenCalledWith(
+          expect.objectContaining({
+            frequencyPenalty
+          })
+        )
+      }
+    })
+
+    it('should support presence penalty', async () => {
+      const penalties = [-2.0, -1.0, 0, 0.5, 1.0, 1.5, 2.0]
+
+      for (const presencePenalty of penalties) {
+        vi.clearAllMocks()
+
+        const mockStream = {
+          textStream: (async function* () {
+            yield 'Response'
+          })(),
+          fullStream: (async function* () {
+            yield { type: 'text-delta', textDelta: 'Response' }
+          })()
+        }
+
+        vi.mocked(streamText).mockResolvedValue(mockStream as any)
+
+        await executor.streamText({
+          model: 'gpt-4',
+          messages: testMessages.simple,
+          presencePenalty
+        })
+
+        expect(streamText).toHaveBeenCalledWith(
+          expect.objectContaining({
+            presencePenalty
+          })
+        )
+      }
+    })
+
+    it('should support both penalties together', async () => {
+      const mockStream = {
+        textStream: (async function* () {
+          yield 'Response'
+        })(),
+        fullStream: (async function* () {
+          yield { type: 'text-delta', textDelta: 'Response' }
+        })()
+      }
+
+      vi.mocked(streamText).mockResolvedValue(mockStream as any)
+
+      await executor.streamText({
+        model: 'gpt-4',
+        messages: testMessages.simple,
+        frequencyPenalty: 0.5,
+        presencePenalty: 0.5
+      })
+
+      expect(streamText).toHaveBeenCalledWith(
+        expect.objectContaining({
+          frequencyPenalty: 0.5,
+          presencePenalty: 0.5
+        })
+      )
+    })
+  })
+
+  describe('Seed Parameter', () => {
+    it('should support seed for deterministic output', async () => {
+      const seeds = [0, 12345, 67890, 999999]
+
+      for (const seed of seeds) {
+        vi.clearAllMocks()
+
+        const mockStream = {
+          textStream: (async function* () {
+            yield 'Response'
+          })(),
+          fullStream: (async function* () {
+            yield { type: 'text-delta', textDelta: 'Response' }
+          })()
+        }
+
+        vi.mocked(streamText).mockResolvedValue(mockStream as any)
+
+        await executor.streamText({
+          model: 'gpt-4',
+          messages: testMessages.simple,
+          seed
+        })
+
+        expect(streamText).toHaveBeenCalledWith(
+          expect.objectContaining({
+            seed
+          })
+        )
+      }
+    })
+  })
+
+  describe('Abort Signal', () => {
+    it('should support abort signal', async () => {
+      const abortController = new AbortController()
+
+      const mockStream = {
+        textStream: (async function* () {
+          yield 'Response'
+        })(),
+        fullStream: (async function* () {
+          yield { type: 'text-delta', textDelta: 'Response' }
+        })()
+      }
+
+      vi.mocked(streamText).mockResolvedValue(mockStream as any)
+
+      await executor.streamText({
+        model: 'gpt-4',
+        messages: testMessages.simple,
+        abortSignal: abortController.signal
+      })
+
+      expect(streamText).toHaveBeenCalledWith(
+        expect.objectContaining({
+          abortSignal: abortController.signal
+        })
+      )
+    })
+
+    it('should handle abort during streaming', async () => {
+      const abortController = new AbortController()
+
+      const mockStream = {
+        textStream: (async function* () {
+          yield 'Start'
+          // Simulate abort
+          abortController.abort()
+          throw new Error('Aborted')
+        })(),
+        fullStream: (async function* () {
+          yield { type: 'text-delta', textDelta: 'Start' }
+          throw new Error('Aborted')
+        })()
+      }
+
+      vi.mocked(streamText).mockResolvedValue(mockStream as any)
+
+      const result = await executor.streamText({
+        model: 'gpt-4',
+        messages: testMessages.simple,
+        abortSignal: abortController.signal
+      })
+
+      await expect(async () => {
+        // oxlint-disable-next-line no-unused-vars
+        for await (const _chunk of result.textStream) {
+          // Stream should be interrupted
+        }
+      }).rejects.toThrow('Aborted')
+    })
+  })
+
+  describe('Plugin Integration', () => {
+    it('should execute plugins during streaming', async () => {
+      const pluginCalls: string[] = []
+
+      const testPlugin: AiPlugin = {
+        name: 'test-plugin',
+        onRequestStart: vi.fn(async () => {
+          pluginCalls.push('onRequestStart')
+        }),
+        transformParams: vi.fn(async (params) => {
+          pluginCalls.push('transformParams')
+          return { ...params, temperature: 0.5 }
+        }),
+        onRequestEnd: vi.fn(async () => {
+          pluginCalls.push('onRequestEnd')
+        })
+      }
+
+      const executorWithPlugin = RuntimeExecutor.create('openai', mockProviderConfigs.openai, [testPlugin])
+
+      const mockStream = {
+        textStream: (async function* () {
+          yield 'Response'
+        })(),
+        fullStream: (async function* () {
+          yield { type: 'text-delta', textDelta: 'Response' }
+        })()
+      }
+
+      vi.mocked(streamText).mockResolvedValue(mockStream as any)
+
+      const result = await executorWithPlugin.streamText({
+        model: 'gpt-4',
+        messages: testMessages.simple
+      })
+
+      // Consume stream
+      // oxlint-disable-next-line no-unused-vars
+      for await (const _chunk of result.textStream) {
+        // Stream chunks
+      }
+
+      expect(pluginCalls).toContain('onRequestStart')
+      expect(pluginCalls).toContain('transformParams')
+
+      // Verify transformed parameters were used
+      expect(streamText).toHaveBeenCalledWith(
+        expect.objectContaining({
+          temperature: 0.5
+        })
+      )
+    })
+  })
+
+  describe('Full Stream with Finish Reason', () => {
+    it('should provide finish reason in full stream', async () => {
+      const mockStream = {
+        textStream: (async function* () {
+          yield 'Response'
+        })(),
+        fullStream: (async function* () {
+          yield { type: 'text-delta', textDelta: 'Response' }
+          yield {
+            type: 'finish',
+            finishReason: 'stop',
+            usage: { promptTokens: 5, completionTokens: 3, totalTokens: 8 }
+          }
+        })()
+      }
+
+      vi.mocked(streamText).mockResolvedValue(mockStream as any)
+
+      const result = await executor.streamText({
+        model: 'gpt-4',
+        messages: testMessages.simple
+      })
+
+      const fullChunks = await collectStreamChunks(result.fullStream)
+
+      expect(fullChunks).toHaveLength(2)
+      expect(fullChunks[0]).toEqual({ type: 'text-delta', textDelta: 'Response' })
+      expect(fullChunks[1]).toEqual({
+        type: 'finish',
+        finishReason: 'stop',
+        usage: { promptTokens: 5, completionTokens: 3, totalTokens: 8 }
+      })
+    })
+  })
+
+  describe('Error Handling', () => {
+    it('should handle streaming errors', async () => {
+      const error = new Error('Streaming failed')
+      vi.mocked(streamText).mockRejectedValue(error)
+
+      await expect(
+        executor.streamText({
+          model: 'gpt-4',
+          messages: testMessages.simple
+        })
+      ).rejects.toThrow('Streaming failed')
+    })
+
+    it('should execute onError plugin hook on failure', async () => {
+      const error = new Error('Stream error')
+      vi.mocked(streamText).mockRejectedValue(error)
+
+      const errorPlugin: AiPlugin = {
+        name: 'error-handler',
+        onError: vi.fn()
+      }
+
+      const executorWithPlugin = RuntimeExecutor.create('openai', mockProviderConfigs.openai, [errorPlugin])
+
+      await expect(
+        executorWithPlugin.streamText({
+          model: 'gpt-4',
+          messages: testMessages.simple
+        })
+      ).rejects.toThrow('Stream error')
+
+      expect(errorPlugin.onError).toHaveBeenCalledWith(
+        error,
+        expect.objectContaining({
+          providerId: 'openai',
+          modelId: 'gpt-4'
+        })
+      )
+    })
+  })
+})
@@ -41,6 +41,7 @@ export enum IpcChannel {
  App_SetFullScreen = 'app:set-full-screen',
  App_IsFullScreen = 'app:is-full-screen',
  App_GetSystemFonts = 'app:get-system-fonts',
+  APP_CrashRenderProcess = 'app:crash-render-process',

  App_MacIsProcessTrusted = 'app:mac-is-process-trusted',
  App_MacRequestProcessTrust = 'app:mac-request-process-trust',
@@ -234,6 +235,7 @@ export enum IpcChannel {
  System_GetDeviceType = 'system:getDeviceType',
  System_GetHostname = 'system:getHostname',
  System_GetCpuName = 'system:getCpuName',
+  System_CheckGitBash = 'system:checkGitBash',

  // DevTools
  System_ToggleDevTools = 'system:toggleDevTools',
@@ -199,7 +199,7 @@ export enum FeedUrl {

 export enum UpdateConfigUrl {
  GITHUB = 'https://raw.githubusercontent.com/CherryHQ/cherry-studio/refs/heads/x-files/app-upgrade-config/app-upgrade-config.json',
-  GITCODE = 'https://raw.gitcode.com/CherryHQ/cherry-studio/raw/x-files/app-upgrade-config/app-upgrade-config.json'
+  GITCODE = 'https://raw.gitcode.com/CherryHQ/cherry-studio/raw/x-files%2Fapp-upgrade-config/app-upgrade-config.json'
 }

 export enum UpgradeChannel {
@@ -4,3 +4,34 @@ export const defaultAppHeaders = () => {
    'X-Title': 'Cherry Studio'
  }
 }
+
+// Following two function are not being used for now.
+// I may use them in the future, so just keep them commented. - by eurfelux
+
+/**
+ * Converts an `undefined` value to `null`, otherwise returns the value as-is.
+ * @param value - The value to check
+ * @returns `null` if the input is `undefined`; otherwise the input value
+ */
+
+// export function toNullIfUndefined<T>(value: T | undefined): T | null {
+//   if (value === undefined) {
+//     return null
+//   } else {
+//     return value
+//   }
+// }
+
+/**
+ * Converts a `null` value to `undefined`, otherwise returns the value as-is.
+ * @param value - The value to check
+ * @returns `undefined` if the input is `null`; otherwise the input value
+ */
+
+// export function toUndefinedIfNull<T>(value: T | null): T | undefined {
+//   if (value === null) {
+//     return undefined
+//   } else {
+//     return value
+//   }
+// }
@@ -104,12 +104,6 @@ const router = express
        logger.warn('No models available from providers', { filter })
      }

-      logger.info('Models response ready', {
-        filter,
-        total: response.total,
-        modelIds: response.data.map((m) => m.id)
-      })
-
      return res.json(response satisfies ApiModelsResponse)
    } catch (error: any) {
      logger.error('Error fetching models', { error })
@@ -3,7 +3,6 @@ import { createServer } from 'node:http'
 import { loggerService } from '@logger'
 import { IpcChannel } from '@shared/IpcChannel'

-import { agentService } from '../services/agents'
 import { windowService } from '../services/WindowService'
 import { app } from './app'
 import { config } from './config'
@@ -32,11 +31,6 @@ export class ApiServer {
    // Load config
    const { port, host } = await config.load()

-    // Initialize AgentService
-    logger.info('Initializing AgentService')
-    await agentService.initialize()
-    logger.info('AgentService initialized')
-
    // Create server with Express app
    this.server = createServer(app)
    this.applyServerTimeouts(this.server)
@@ -32,7 +32,7 @@ export class ModelsService {

      for (const model of models) {
        const provider = providers.find((p) => p.id === model.provider)
-        logger.debug(`Processing model ${model.id}`)
+        // logger.debug(`Processing model ${model.id}`)
        if (!provider) {
          logger.debug(`Skipping model ${model.id} . Reason: Provider not found.`)
          continue
@@ -8,7 +8,7 @@ import '@main/config'
 import { loggerService } from '@logger'
 import { electronApp, optimizer } from '@electron-toolkit/utils'
 import { replaceDevtoolsFont } from '@main/utils/windowUtil'
-import { app } from 'electron'
+import { app, crashReporter } from 'electron'
 import installExtension, { REACT_DEVELOPER_TOOLS, REDUX_DEVTOOLS } from 'electron-devtools-installer'
 import { isDev, isLinux, isWin } from './constant'

@@ -34,9 +34,18 @@ import { TrayService } from './services/TrayService'
 import { versionService } from './services/VersionService'
 import { windowService } from './services/WindowService'
 import { initWebviewHotkeys } from './services/WebviewService'
+import { runAsyncFunction } from './utils'

 const logger = loggerService.withContext('MainEntry')

+// enable local crash reports
+crashReporter.start({
+  companyName: 'CherryHQ',
+  productName: 'CherryStudio',
+  submitURL: '',
+  uploadToServer: false
+})
+
 /**
 * Disable hardware acceleration if setting is enabled
 */
@@ -162,39 +171,33 @@ if (!app.requestSingleInstanceLock()) {
    //start selection assistant service
    initSelectionService()

-    // Initialize Agent Service
-    try {
-      await agentService.initialize()
-      logger.info('Agent service initialized successfully')
-    } catch (error: any) {
-      logger.error('Failed to initialize Agent service:', error)
-    }
+    runAsyncFunction(async () => {
+      // Start API server if enabled or if agents exist
+      try {
+        const config = await apiServerService.getCurrentConfig()
+        logger.info('API server config:', config)

-    // Start API server if enabled or if agents exist
-    try {
-      const config = await apiServerService.getCurrentConfig()
-      logger.info('API server config:', config)
-
-      // Check if there are any agents
-      let shouldStart = config.enabled
-      if (!shouldStart) {
-        try {
-          const { total } = await agentService.listAgents({ limit: 1 })
-          if (total > 0) {
-            shouldStart = true
-            logger.info(`Detected ${total} agent(s), auto-starting API server`)
+        // Check if there are any agents
+        let shouldStart = config.enabled
+        if (!shouldStart) {
+          try {
+            const { total } = await agentService.listAgents({ limit: 1 })
+            if (total > 0) {
+              shouldStart = true
+              logger.info(`Detected ${total} agent(s), auto-starting API server`)
+            }
+          } catch (error: any) {
+            logger.warn('Failed to check agent count:', error)
          }
-        } catch (error: any) {
-          logger.warn('Failed to check agent count:', error)
        }
-      }

-      if (shouldStart) {
-        await apiServerService.start()
+        if (shouldStart) {
+          await apiServerService.start()
+        }
+      } catch (error: any) {
+        logger.error('Failed to check/start API server:', error)
      }
-    } catch (error: any) {
-      logger.error('Failed to check/start API server:', error)
-    }
+    })
  })

  registerProtocolClient(app)
@@ -493,6 +493,44 @@ export function registerIpc(mainWindow: BrowserWindow, app: Electron.App) {
  ipcMain.handle(IpcChannel.System_GetDeviceType, () => (isMac ? 'mac' : isWin ? 'windows' : 'linux'))
  ipcMain.handle(IpcChannel.System_GetHostname, () => require('os').hostname())
  ipcMain.handle(IpcChannel.System_GetCpuName, () => require('os').cpus()[0].model)
+  ipcMain.handle(IpcChannel.System_CheckGitBash, () => {
+    if (!isWin) {
+      return true // Non-Windows systems don't need Git Bash
+    }
+
+    try {
+      // Check common Git Bash installation paths
+      const commonPaths = [
+        path.join(process.env.ProgramFiles || 'C:\\Program Files', 'Git', 'bin', 'bash.exe'),
+        path.join(process.env['ProgramFiles(x86)'] || 'C:\\Program Files (x86)', 'Git', 'bin', 'bash.exe'),
+        path.join(process.env.LOCALAPPDATA || '', 'Programs', 'Git', 'bin', 'bash.exe')
+      ]
+
+      // Check if any of the common paths exist
+      for (const bashPath of commonPaths) {
+        if (fs.existsSync(bashPath)) {
+          logger.debug('Git Bash found', { path: bashPath })
+          return true
+        }
+      }
+
+      // Check if git is in PATH
+      const { execSync } = require('child_process')
+      try {
+        execSync('git --version', { stdio: 'ignore' })
+        logger.debug('Git found in PATH')
+        return true
+      } catch {
+        // Git not in PATH
+      }
+
+      logger.debug('Git Bash not found on Windows system')
+      return false
+    } catch (error) {
+      logger.error('Error checking Git Bash', error as Error)
+      return false
+    }
+  })
  ipcMain.handle(IpcChannel.System_ToggleDevTools, (e) => {
    const win = BrowserWindow.fromWebContents(e.sender)
    win && win.webContents.toggleDevTools()
@@ -1038,4 +1076,8 @@ export function registerIpc(mainWindow: BrowserWindow, app: Electron.App) {
  ipcMain.handle(IpcChannel.WebSocket_Status, WebSocketService.getStatus)
  ipcMain.handle(IpcChannel.WebSocket_SendFile, WebSocketService.sendFile)
  ipcMain.handle(IpcChannel.WebSocket_GetAllCandidates, WebSocketService.getAllCandidates)
+
+  ipcMain.handle(IpcChannel.APP_CrashRenderProcess, () => {
+    mainWindow.webContents.forcefullyCrashRenderer()
+  })
 }
@@ -21,6 +21,7 @@ type ApiResponse<T> = {
 type BatchUploadResponse = {
  batch_id: string
  file_urls: string[]
+  headers?: Record<string, string>[]
 }

 type ExtractProgress = {
@@ -55,7 +56,7 @@ type QuotaResponse = {
 export default class MineruPreprocessProvider extends BasePreprocessProvider {
  constructor(provider: PreprocessProvider, userId?: string) {
    super(provider, userId)
-    // todo：免费期结束后删除
+    // TODO: remove after free period ends
    this.provider.apiKey = this.provider.apiKey || import.meta.env.MAIN_VITE_MINERU_API_KEY
  }

@@ -68,21 +69,21 @@ export default class MineruPreprocessProvider extends BasePreprocessProvider {
      logger.info(`MinerU preprocess processing started: ${filePath}`)
      await this.validateFile(filePath)

-      // 1. 获取上传URL并上传文件
+      // 1. Get upload URL and upload file
      const batchId = await this.uploadFile(file)
      logger.info(`MinerU file upload completed: batch_id=${batchId}`)

-      // 2. 等待处理完成并获取结果
+      // 2. Wait for completion and fetch results
      const extractResult = await this.waitForCompletion(sourceId, batchId, file.origin_name)
      logger.info(`MinerU processing completed for batch: ${batchId}`)

-      // 3. 下载并解压文件
+      // 3. Download and extract output
      const { path: outputPath } = await this.downloadAndExtractFile(extractResult.full_zip_url!, file)

      // 4. check quota
      const quota = await this.checkQuota()

-      // 5. 创建处理后的文件信息
+      // 5. Create processed file metadata
      return {
        processedFile: this.createProcessedFileInfo(file, outputPath),
        quota
@@ -115,23 +116,48 @@ export default class MineruPreprocessProvider extends BasePreprocessProvider {
  }

  private async validateFile(filePath: string): Promise<void> {
+    // Phase 1: check file size (without loading into memory)
+    logger.info(`Validating PDF file: ${filePath}`)
+    const stats = await fs.promises.stat(filePath)
+    const fileSizeBytes = stats.size
+
+    // Ensure file size is under 200MB
+    if (fileSizeBytes >= 200 * 1024 * 1024) {
+      const fileSizeMB = Math.round(fileSizeBytes / (1024 * 1024))
+      throw new Error(`PDF file size (${fileSizeMB}MB) exceeds the limit of 200MB`)
+    }
+
+    // Phase 2: check page count (requires reading file with error handling)
    const pdfBuffer = await fs.promises.readFile(filePath)

-    const doc = await this.readPdf(pdfBuffer)
+    try {
+      const doc = await this.readPdf(pdfBuffer)

-    // 文件页数小于600页
-    if (doc.numPages >= 600) {
-      throw new Error(`PDF page count (${doc.numPages}) exceeds the limit of 600 pages`)
-    }
-    // 文件大小小于200MB
-    if (pdfBuffer.length >= 200 * 1024 * 1024) {
-      const fileSizeMB = Math.round(pdfBuffer.length / (1024 * 1024))
-      throw new Error(`PDF file size (${fileSizeMB}MB) exceeds the limit of 200MB`)
+      // Ensure page count is under 600 pages
+      if (doc.numPages >= 600) {
+        throw new Error(`PDF page count (${doc.numPages}) exceeds the limit of 600 pages`)
+      }
+
+      logger.info(`PDF validation passed: ${doc.numPages} pages, ${Math.round(fileSizeBytes / (1024 * 1024))}MB`)
+    } catch (error: any) {
+      // If the page limit is exceeded, rethrow immediately
+      if (error.message.includes('exceeds the limit')) {
+        throw error
+      }
+
+      // If PDF parsing fails, log a detailed warning but continue processing
+      logger.warn(
+        `Failed to parse PDF structure (file may be corrupted or use non-standard format). ` +
+          `Skipping page count validation. Will attempt to process with MinerU API. ` +
+          `Error details: ${error.message}. ` +
+          `Suggestion: If processing fails, try repairing the PDF using tools like Adobe Acrobat or online PDF repair services.`
+      )
+      // Do not throw; continue processing
    }
  }

  private createProcessedFileInfo(file: FileMetadata, outputPath: string): FileMetadata {
-    // 查找解压后的主要文件
+    // Locate the main extracted file
    let finalPath = ''
    let finalName = file.origin_name.replace('.pdf', '.md')

@@ -143,14 +169,14 @@ export default class MineruPreprocessProvider extends BasePreprocessProvider {
        const originalMdPath = path.join(outputPath, mdFile)
        const newMdPath = path.join(outputPath, finalName)

-        // 重命名文件为原始文件名
+        // Rename the file to match the original name
        try {
          fs.renameSync(originalMdPath, newMdPath)
          finalPath = newMdPath
          logger.info(`Renamed markdown file from ${mdFile} to ${finalName}`)
        } catch (renameError) {
          logger.warn(`Failed to rename file ${mdFile} to ${finalName}: ${renameError}`)
-          // 如果重命名失败，使用原文件
+          // If renaming fails, fall back to the original file
          finalPath = originalMdPath
          finalName = mdFile
        }
@@ -178,7 +204,7 @@ export default class MineruPreprocessProvider extends BasePreprocessProvider {
    logger.info(`Downloading MinerU result to: ${zipPath}`)

    try {
-      // 下载ZIP文件
+      // Download the ZIP file
      const response = await net.fetch(zipUrl, { method: 'GET' })
      if (!response.ok) {
        throw new Error(`HTTP ${response.status}: ${response.statusText}`)
@@ -187,17 +213,17 @@ export default class MineruPreprocessProvider extends BasePreprocessProvider {
      fs.writeFileSync(zipPath, Buffer.from(arrayBuffer))
      logger.info(`Downloaded ZIP file: ${zipPath}`)

-      // 确保提取目录存在
+      // Ensure the extraction directory exists
      if (!fs.existsSync(extractPath)) {
        fs.mkdirSync(extractPath, { recursive: true })
      }

-      // 解压文件
+      // Extract the ZIP contents
      const zip = new AdmZip(zipPath)
      zip.extractAllTo(extractPath, true)
      logger.info(`Extracted files to: ${extractPath}`)

-      // 删除临时ZIP文件
+      // Remove the temporary ZIP file
      fs.unlinkSync(zipPath)

      return { path: extractPath }
@@ -209,11 +235,11 @@ export default class MineruPreprocessProvider extends BasePreprocessProvider {

  private async uploadFile(file: FileMetadata): Promise<string> {
    try {
-      // 步骤1: 获取上传URL
-      const { batchId, fileUrls } = await this.getBatchUploadUrls(file)
-      // 步骤2: 上传文件到获取的URL
+      // Step 1: obtain the upload URL
+      const { batchId, fileUrls, uploadHeaders } = await this.getBatchUploadUrls(file)
+      // Step 2: upload the file to the obtained URL
      const filePath = fileStorage.getFilePathById(file)
-      await this.putFileToUrl(filePath, fileUrls[0])
+      await this.putFileToUrl(filePath, fileUrls[0], file.origin_name, uploadHeaders?.[0])
      logger.info(`File uploaded successfully: ${filePath}`, { batchId, fileUrls })

      return batchId
@@ -223,7 +249,9 @@ export default class MineruPreprocessProvider extends BasePreprocessProvider {
    }
  }

-  private async getBatchUploadUrls(file: FileMetadata): Promise<{ batchId: string; fileUrls: string[] }> {
+  private async getBatchUploadUrls(
+    file: FileMetadata
+  ): Promise<{ batchId: string; fileUrls: string[]; uploadHeaders?: Record<string, string>[] }> {
    const endpoint = `${this.provider.apiHost}/api/v4/file-urls/batch`

    const payload = {
@@ -254,10 +282,11 @@ export default class MineruPreprocessProvider extends BasePreprocessProvider {
      if (response.ok) {
        const data: ApiResponse<BatchUploadResponse> = await response.json()
        if (data.code === 0 && data.data) {
-          const { batch_id, file_urls } = data.data
+          const { batch_id, file_urls, headers: uploadHeaders } = data.data
          return {
            batchId: batch_id,
-            fileUrls: file_urls
+            fileUrls: file_urls,
+            uploadHeaders
          }
        } else {
          throw new Error(`API returned error: ${data.msg || JSON.stringify(data)}`)
@@ -271,18 +300,28 @@ export default class MineruPreprocessProvider extends BasePreprocessProvider {
    }
  }

-  private async putFileToUrl(filePath: string, uploadUrl: string): Promise<void> {
+  private async putFileToUrl(
+    filePath: string,
+    uploadUrl: string,
+    fileName?: string,
+    headers?: Record<string, string>
+  ): Promise<void> {
    try {
      const fileBuffer = await fs.promises.readFile(filePath)
+      const fileSize = fileBuffer.byteLength
+      const displayName = fileName ?? path.basename(filePath)
+
+      logger.info(`Uploading file to MinerU OSS: ${displayName} (${fileSize} bytes)`)

      // https://mineru.net/apiManage/docs
      const response = await net.fetch(uploadUrl, {
        method: 'PUT',
-        body: fileBuffer
+        headers,
+        body: new Uint8Array(fileBuffer)
      })

      if (!response.ok) {
-        // 克隆 response 以避免消费 body stream
+        // Clone the response to avoid consuming the body stream
        const responseClone = response.clone()

        try {
@@ -353,20 +392,20 @@ export default class MineruPreprocessProvider extends BasePreprocessProvider {
      try {
        const result = await this.getExtractResults(batchId)

-        // 查找对应文件的处理结果
+        // Find the corresponding file result
        const fileResult = result.extract_result.find((item) => item.file_name === fileName)
        if (!fileResult) {
          throw new Error(`File ${fileName} not found in batch results`)
        }

-        // 检查处理状态
+        // Check the processing state
        if (fileResult.state === 'done' && fileResult.full_zip_url) {
          logger.info(`Processing completed for file: ${fileName}`)
          return fileResult
        } else if (fileResult.state === 'failed') {
          throw new Error(`Processing failed for file: ${fileName}, error: ${fileResult.err_msg}`)
        } else if (fileResult.state === 'running') {
-          // 发送进度更新
+          // Send progress updates
          if (fileResult.extract_progress) {
            const progress = Math.round(
              (fileResult.extract_progress.extracted_pages / fileResult.extract_progress.total_pages) * 100
@@ -374,7 +413,7 @@ export default class MineruPreprocessProvider extends BasePreprocessProvider {
            await this.sendPreprocessProgress(sourceId, progress)
            logger.info(`File ${fileName} processing progress: ${progress}%`)
          } else {
-            // 如果没有具体进度信息，发送一个通用进度
+            // If no detailed progress information is available, send a generic update
            await this.sendPreprocessProgress(sourceId, 50)
            logger.info(`File ${fileName} is still processing...`)
          }
@@ -53,18 +53,43 @@ export default class OpenMineruPreprocessProvider extends BasePreprocessProvider
  }

  private async validateFile(filePath: string): Promise<void> {
+    // 第一阶段:检查文件大小(无需读取文件到内存)
+    logger.info(`Validating PDF file: ${filePath}`)
+    const stats = await fs.promises.stat(filePath)
+    const fileSizeBytes = stats.size
+
+    // File size must be less than 200MB
+    if (fileSizeBytes >= 200 * 1024 * 1024) {
+      const fileSizeMB = Math.round(fileSizeBytes / (1024 * 1024))
+      throw new Error(`PDF file size (${fileSizeMB}MB) exceeds the limit of 200MB`)
+    }
+
+    // 第二阶段:检查页数(需要读取文件,带错误处理)
    const pdfBuffer = await fs.promises.readFile(filePath)

-    const doc = await this.readPdf(pdfBuffer)
+    try {
+      const doc = await this.readPdf(pdfBuffer)

-    // File page count must be less than 600 pages
-    if (doc.numPages >= 600) {
-      throw new Error(`PDF page count (${doc.numPages}) exceeds the limit of 600 pages`)
-    }
-    // File size must be less than 200MB
-    if (pdfBuffer.length >= 200 * 1024 * 1024) {
-      const fileSizeMB = Math.round(pdfBuffer.length / (1024 * 1024))
-      throw new Error(`PDF file size (${fileSizeMB}MB) exceeds the limit of 200MB`)
+      // File page count must be less than 600 pages
+      if (doc.numPages >= 600) {
+        throw new Error(`PDF page count (${doc.numPages}) exceeds the limit of 600 pages`)
+      }
+
+      logger.info(`PDF validation passed: ${doc.numPages} pages, ${Math.round(fileSizeBytes / (1024 * 1024))}MB`)
+    } catch (error: any) {
+      // 如果是页数超限错误,直接抛出
+      if (error.message.includes('exceeds the limit')) {
+        throw error
+      }
+
+      // PDF 解析失败,记录详细警告但允许继续处理
+      logger.warn(
+        `Failed to parse PDF structure (file may be corrupted or use non-standard format). ` +
+          `Skipping page count validation. Will attempt to process with MinerU API. ` +
+          `Error details: ${error.message}. ` +
+          `Suggestion: If processing fails, try repairing the PDF using tools like Adobe Acrobat or online PDF repair services.`
+      )
+      // 不抛出错误,允许继续处理
    }
  }

@@ -72,8 +97,8 @@ export default class OpenMineruPreprocessProvider extends BasePreprocessProvider
    // Find the main file after extraction
    let finalPath = ''
    let finalName = file.origin_name.replace('.pdf', '.md')
-    // Find the corresponding folder by file name
-    outputPath = path.join(outputPath, `${file.origin_name.replace('.pdf', '')}`)
+    // Find the corresponding folder by file id
+    outputPath = path.join(outputPath, file.id)
    try {
      const files = fs.readdirSync(outputPath)

@@ -125,7 +150,7 @@ export default class OpenMineruPreprocessProvider extends BasePreprocessProvider
    formData.append('return_md', 'true')
    formData.append('response_format_zip', 'true')
    formData.append('files', fileBuffer, {
-      filename: file.origin_name
+      filename: file.name
    })

    while (retries < maxRetries) {
@@ -139,7 +164,7 @@ export default class OpenMineruPreprocessProvider extends BasePreprocessProvider
            ...(this.provider.apiKey ? { Authorization: `Bearer ${this.provider.apiKey}` } : {}),
            ...formData.getHeaders()
          },
-          body: formData.getBuffer()
+          body: new Uint8Array(formData.getBuffer())
        })

        if (!response.ok) {
@@ -8,6 +8,7 @@ import DiDiMcpServer from './didi-mcp'
 import DifyKnowledgeServer from './dify-knowledge'
 import FetchServer from './fetch'
 import FileSystemServer from './filesystem'
+import MCPUIDemoServer from './mcp-ui-demo'
 import MemoryServer from './memory'
 import PythonServer from './python'
 import ThinkingServer from './sequentialthinking'
@@ -48,6 +49,9 @@ export function createInMemoryMCPServer(
      const apiKey = envs.DIDI_API_KEY
      return new DiDiMcpServer(apiKey).server
    }
+    case BuiltinMCPServerNames.mcpUIDemo: {
+      return new MCPUIDemoServer().server
+    }
    default:
      throw new Error(`Unknown in-memory MCP server: ${name}`)
  }
@@ -0,0 +1,433 @@
+import { Server } from '@modelcontextprotocol/sdk/server/index.js'
+import { CallToolRequestSchema, ListToolsRequestSchema } from '@modelcontextprotocol/sdk/types.js'
+
+const server = new Server(
+  {
+    name: 'mcp-ui-demo',
+    version: '1.0.0'
+  },
+  {
+    capabilities: {
+      tools: {}
+    }
+  }
+)
+
+// HTML templates for different UIs
+const getHelloWorldUI = () =>
+  `
+<!DOCTYPE html>
+<html>
+<head>
+  <style>
+    body {
+      font-family: system-ui, -apple-system, sans-serif;
+      padding: 20px;
+      background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
+      color: white;
+      min-height: 200px;
+      display: flex;
+      align-items: center;
+      justify-content: center;
+      margin: 0;
+    }
+    .container {
+      text-align: center;
+    }
+    h1 {
+      font-size: 2.5em;
+      margin-bottom: 10px;
+      text-shadow: 2px 2px 4px rgba(0,0,0,0.3);
+    }
+    p {
+      font-size: 1.2em;
+      opacity: 0.9;
+    }
+  </style>
+</head>
+<body>
+  <div class="container">
+    <h1>🎉 Hello from MCP UI!</h1>
+    <p>This is a simple MCP UI Resource rendered in Cherry Studio</p>
+  </div>
+</body>
+</html>
+`.trim()
+
+const getInteractiveUI = () =>
+  `
+<!DOCTYPE html>
+<html>
+<head>
+  <style>
+    body {
+      font-family: system-ui, -apple-system, sans-serif;
+      padding: 20px;
+      background: #f5f5f5;
+      margin: 0;
+    }
+    .container {
+      max-width: 600px;
+      margin: 0 auto;
+      background: white;
+      padding: 30px;
+      border-radius: 12px;
+      box-shadow: 0 2px 10px rgba(0,0,0,0.1);
+    }
+    h2 {
+      color: #333;
+      margin-bottom: 20px;
+    }
+    button {
+      background: #667eea;
+      color: white;
+      border: none;
+      padding: 12px 24px;
+      border-radius: 6px;
+      cursor: pointer;
+      font-size: 16px;
+      margin: 5px;
+      transition: background 0.2s;
+    }
+    button:hover {
+      background: #5568d3;
+    }
+    #output {
+      margin-top: 20px;
+      padding: 15px;
+      background: #f8f9fa;
+      border-radius: 6px;
+      border: 1px solid #e0e0e0;
+      min-height: 50px;
+    }
+    .info {
+      color: #666;
+      font-size: 14px;
+      margin-top: 15px;
+    }
+  </style>
+</head>
+<body>
+  <div class="container">
+    <h2>Interactive MCP UI Demo</h2>
+    <p>Click the buttons to interact with MCP tools:</p>
+
+    <button onclick="callEchoTool()">Call Echo Tool</button>
+    <button onclick="getTimestamp()">Get Timestamp</button>
+    <button onclick="openLink()">Open External Link</button>
+
+    <div id="output"></div>
+
+    <div class="info">
+      This UI can communicate with the host application through postMessage API.
+    </div>
+  </div>
+
+  <script>
+    function callEchoTool() {
+      const output = document.getElementById('output');
+      output.innerHTML = '<p style="color: #0066cc;">Calling echo tool...</p>';
+
+      window.parent.postMessage({
+        type: 'tool',
+        payload: {
+          toolName: 'demo_echo',
+          params: {
+            message: 'Hello from MCP UI! Time: ' + new Date().toLocaleTimeString()
+          }
+        }
+      }, '*');
+    }
+
+    function getTimestamp() {
+      const output = document.getElementById('output');
+      const now = new Date();
+      output.innerHTML = \`
+        <p style="color: #00aa00;">
+          <strong>Current Timestamp:</strong><br/>
+          \${now.toLocaleString()}<br/>
+          Unix: \${Math.floor(now.getTime() / 1000)}
+        </p>
+      \`;
+    }
+
+    function openLink() {
+      window.parent.postMessage({
+        type: 'link',
+        payload: {
+          url: 'https://github.com/idosal/mcp-ui'
+        }
+      }, '*');
+    }
+
+    // Listen for responses
+    window.addEventListener('message', (event) => {
+      if (event.data.type === 'ui-message-response') {
+        const output = document.getElementById('output');
+        const { response, error } = event.data.payload;
+
+        if (error) {
+          output.innerHTML = \`<p style="color: #cc0000;">Error: \${error}</p>\`;
+        } else {
+          output.innerHTML = \`<p style="color: #00aa00;">Response: \${JSON.stringify(response, null, 2)}</p>\`;
+        }
+      }
+    });
+  </script>
+</body>
+</html>
+`.trim()
+
+const getFormUI = () =>
+  `
+<!DOCTYPE html>
+<html>
+<head>
+  <style>
+    body {
+      font-family: system-ui, -apple-system, sans-serif;
+      padding: 20px;
+      background: #f5f5f5;
+      margin: 0;
+    }
+    .container {
+      max-width: 500px;
+      margin: 0 auto;
+      background: white;
+      padding: 30px;
+      border-radius: 12px;
+      box-shadow: 0 2px 10px rgba(0,0,0,0.1);
+    }
+    h2 {
+      color: #333;
+      margin-bottom: 20px;
+    }
+    .form-group {
+      margin-bottom: 15px;
+    }
+    label {
+      display: block;
+      margin-bottom: 5px;
+      color: #555;
+      font-weight: 500;
+    }
+    input, textarea {
+      width: 100%;
+      padding: 10px;
+      border: 1px solid #ddd;
+      border-radius: 6px;
+      font-size: 14px;
+      box-sizing: border-box;
+    }
+    textarea {
+      min-height: 100px;
+      resize: vertical;
+    }
+    button {
+      background: #667eea;
+      color: white;
+      border: none;
+      padding: 12px 24px;
+      border-radius: 6px;
+      cursor: pointer;
+      font-size: 16px;
+      width: 100%;
+      margin-top: 10px;
+    }
+    button:hover {
+      background: #5568d3;
+    }
+    #result {
+      margin-top: 20px;
+      padding: 15px;
+      background: #f8f9fa;
+      border-radius: 6px;
+      display: none;
+    }
+  </style>
+</head>
+<body>
+  <div class="container">
+    <h2>📝 Form UI Demo</h2>
+    <form id="demoForm" onsubmit="handleSubmit(event)">
+      <div class="form-group">
+        <label for="name">Name:</label>
+        <input type="text" id="name" name="name" required placeholder="Enter your name">
+      </div>
+
+      <div class="form-group">
+        <label for="email">Email:</label>
+        <input type="email" id="email" name="email" required placeholder="your@email.com">
+      </div>
+
+      <div class="form-group">
+        <label for="message">Message:</label>
+        <textarea id="message" name="message" required placeholder="Enter your message here..."></textarea>
+      </div>
+
+      <button type="submit">Submit Form</button>
+    </form>
+
+    <div id="result"></div>
+  </div>
+
+  <script>
+    function handleSubmit(event) {
+      event.preventDefault();
+
+      const formData = new FormData(event.target);
+      const data = Object.fromEntries(formData.entries());
+
+      const result = document.getElementById('result');
+      result.style.display = 'block';
+      result.innerHTML = '<p style="color: #0066cc;">Submitting form...</p>';
+
+      // Send form data to host
+      window.parent.postMessage({
+        type: 'notify',
+        payload: {
+          message: 'Form submitted with data: ' + JSON.stringify(data)
+        }
+      }, '*');
+
+      // Display result
+      result.innerHTML = \`
+        <p style="color: #00aa00;"><strong>Form Submitted!</strong></p>
+        <pre style="background: white; padding: 10px; border-radius: 4px; overflow-x: auto;">\${JSON.stringify(data, null, 2)}</pre>
+      \`;
+    }
+  </script>
+</body>
+</html>
+`.trim()
+
+// List available tools
+server.setRequestHandler(ListToolsRequestSchema, async () => {
+  return {
+    tools: [
+      {
+        name: 'demo_echo',
+        description: 'Echo back the message sent from UI',
+        inputSchema: {
+          type: 'object',
+          properties: {
+            message: {
+              type: 'string',
+              description: 'Message to echo back'
+            }
+          },
+          required: ['message']
+        }
+      },
+      {
+        name: 'show_hello_ui',
+        description: 'Display a simple hello world UI with gradient background',
+        inputSchema: {
+          type: 'object',
+          properties: {}
+        }
+      },
+      {
+        name: 'show_interactive_ui',
+        description:
+          'Display an interactive UI demo with buttons for calling tools, getting timestamps, and opening links',
+        inputSchema: {
+          type: 'object',
+          properties: {}
+        }
+      },
+      {
+        name: 'show_form_ui',
+        description: 'Display a form UI demo with input fields for name, email, and message',
+        inputSchema: {
+          type: 'object',
+          properties: {}
+        }
+      }
+    ]
+  }
+})
+
+// Handle tool calls
+server.setRequestHandler(CallToolRequestSchema, async (request) => {
+  const { name, arguments: args } = request.params
+
+  if (name === 'demo_echo') {
+    return {
+      content: [
+        {
+          type: 'text',
+          text: JSON.stringify({
+            echo: args?.message || 'No message provided',
+            timestamp: new Date().toISOString()
+          })
+        }
+      ]
+    }
+  }
+
+  if (name === 'show_hello_ui') {
+    return {
+      content: [
+        {
+          type: 'text',
+          text: JSON.stringify({
+            type: 'resource',
+            resource: {
+              uri: 'ui://demo/hello',
+              mimeType: 'text/html',
+              text: getHelloWorldUI()
+            }
+          })
+        }
+      ]
+    }
+  }
+
+  if (name === 'show_interactive_ui') {
+    return {
+      content: [
+        {
+          type: 'text',
+          text: JSON.stringify({
+            type: 'resource',
+            resource: {
+              uri: 'ui://demo/interactive',
+              mimeType: 'text/html',
+              text: getInteractiveUI()
+            }
+          })
+        }
+      ]
+    }
+  }
+
+  if (name === 'show_form_ui') {
+    return {
+      content: [
+        {
+          type: 'text',
+          text: JSON.stringify({
+            type: 'resource',
+            resource: {
+              uri: 'ui://demo/form',
+              mimeType: 'text/html',
+              text: getFormUI()
+            }
+          })
+        }
+      ]
+    }
+  }
+
+  throw new Error(`Unknown tool: ${name}`)
+})
+
+class MCPUIDemoServer {
+  public server: Server
+  constructor() {
+    this.server = server
+  }
+}
+
+export default MCPUIDemoServer
@@ -1,17 +1,13 @@
-import { type Client, createClient } from '@libsql/client'
 import { loggerService } from '@logger'
 import { mcpApiService } from '@main/apiServer/services/mcp'
 import type { ModelValidationError } from '@main/apiServer/utils'
 import { validateModelId } from '@main/apiServer/utils'
 import type { AgentType, MCPTool, SlashCommand, Tool } from '@types'
 import { objectKeys } from '@types'
-import { drizzle, type LibSQLDatabase } from 'drizzle-orm/libsql'
 import fs from 'fs'
 import path from 'path'

-import { MigrationService } from './database/MigrationService'
-import * as schema from './database/schema'
-import { dbPath } from './drizzle.config'
+import { DatabaseManager } from './database/DatabaseManager'
 import type { AgentModelField } from './errors'
 import { AgentModelValidationError } from './errors'
 import { builtinSlashCommands } from './services/claudecode/commands'
@@ -20,22 +16,16 @@ import { builtinTools } from './services/claudecode/tools'
 const logger = loggerService.withContext('BaseService')

 /**
- * Base service class providing shared database connection and utilities
- * for all agent-related services.
+ * Base service class providing shared utilities for all agent-related services.
 *
 * Features:
- * - Programmatic schema management (no CLI dependencies)
- * - Automatic table creation and migration
- * - Schema version tracking and compatibility checks
- * - Transaction-based operations for safety
- * - Development vs production mode handling
- * - Connection retry logic with exponential backoff
+ * - Database access through DatabaseManager singleton
+ * - JSON field serialization/deserialization
+ * - Path validation and creation
+ * - Model validation
+ * - MCP tools and slash commands listing
 */
 export abstract class BaseService {
-  protected static client: Client | null = null
-  protected static db: LibSQLDatabase<typeof schema> | null = null
-  protected static isInitialized = false
-  protected static initializationPromise: Promise<void> | null = null
  protected jsonFields: string[] = [
    'tools',
    'mcps',
@@ -45,23 +35,6 @@ export abstract class BaseService {
    'slash_commands'
  ]

-  /**
-   * Initialize database with retry logic and proper error handling
-   */
-  protected static async initialize(): Promise<void> {
-    // Return existing initialization if in progress
-    if (BaseService.initializationPromise) {
-      return BaseService.initializationPromise
-    }
-
-    if (BaseService.isInitialized) {
-      return
-    }
-
-    BaseService.initializationPromise = BaseService.performInitialization()
-    return BaseService.initializationPromise
-  }
-
  public async listMcpTools(agentType: AgentType, ids?: string[]): Promise<Tool[]> {
    const tools: Tool[] = []
    if (agentType === 'claude-code') {
@@ -101,78 +74,13 @@ export abstract class BaseService {
    return []
  }

-  private static async performInitialization(): Promise<void> {
-    const maxRetries = 3
-    let lastError: Error
-
-    for (let attempt = 1; attempt <= maxRetries; attempt++) {
-      try {
-        logger.info(`Initializing Agent database at: ${dbPath} (attempt ${attempt}/${maxRetries})`)
-
-        // Ensure the database directory exists
-        const dbDir = path.dirname(dbPath)
-        if (!fs.existsSync(dbDir)) {
-          logger.info(`Creating database directory: ${dbDir}`)
-          fs.mkdirSync(dbDir, { recursive: true })
-        }
-
-        BaseService.client = createClient({
-          url: `file:${dbPath}`
-        })
-
-        BaseService.db = drizzle(BaseService.client, { schema })
-
-        // Run database migrations
-        const migrationService = new MigrationService(BaseService.db, BaseService.client)
-        await migrationService.runMigrations()
-
-        BaseService.isInitialized = true
-        logger.info('Agent database initialized successfully')
-        return
-      } catch (error) {
-        lastError = error as Error
-        logger.warn(`Database initialization attempt ${attempt} failed:`, lastError)
-
-        // Clean up on failure
-        if (BaseService.client) {
-          try {
-            BaseService.client.close()
-          } catch (closeError) {
-            logger.warn('Failed to close client during cleanup:', closeError as Error)
-          }
-        }
-        BaseService.client = null
-        BaseService.db = null
-
-        // Wait before retrying (exponential backoff)
-        if (attempt < maxRetries) {
-          const delay = Math.pow(2, attempt) * 1000 // 2s, 4s, 8s
-          logger.info(`Retrying in ${delay}ms...`)
-          await new Promise((resolve) => setTimeout(resolve, delay))
-        }
-      }
-    }
-
-    // All retries failed
-    BaseService.initializationPromise = null
-    logger.error('Failed to initialize Agent database after all retries:', lastError!)
-    throw lastError!
-  }
-
-  protected ensureInitialized(): void {
-    if (!BaseService.isInitialized || !BaseService.db || !BaseService.client) {
-      throw new Error('Database not initialized. Call initialize() first.')
-    }
-  }
-
-  protected get database(): LibSQLDatabase<typeof schema> {
-    this.ensureInitialized()
-    return BaseService.db!
-  }
-
-  protected get rawClient(): Client {
-    this.ensureInitialized()
-    return BaseService.client!
+  /**
+   * Get database instance
+   * Automatically waits for initialization to complete
+   */
+  protected async getDatabase() {
+    const dbManager = await DatabaseManager.getInstance()
+    return dbManager.getDatabase()
  }

  protected serializeJsonFields(data: any): any {
@@ -284,7 +192,7 @@ export abstract class BaseService {
  }

  /**
-   * Force re-initialization (for development/testing)
+   * Validate agent model configuration
   */
  protected async validateAgentModels(
    agentType: AgentType,
@@ -325,22 +233,4 @@ export abstract class BaseService {
      }
    }
  }
-
-  static async reinitialize(): Promise<void> {
-    BaseService.isInitialized = false
-    BaseService.initializationPromise = null
-
-    if (BaseService.client) {
-      try {
-        BaseService.client.close()
-      } catch (error) {
-        logger.warn('Failed to close client during reinitialize:', error as Error)
-      }
-    }
-
-    BaseService.client = null
-    BaseService.db = null
-
-    await BaseService.initialize()
-  }
 }
@@ -0,0 +1,156 @@
+import { type Client, createClient } from '@libsql/client'
+import { loggerService } from '@logger'
+import type { LibSQLDatabase } from 'drizzle-orm/libsql'
+import { drizzle } from 'drizzle-orm/libsql'
+import fs from 'fs'
+import path from 'path'
+
+import { dbPath } from '../drizzle.config'
+import { MigrationService } from './MigrationService'
+import * as schema from './schema'
+
+const logger = loggerService.withContext('DatabaseManager')
+
+/**
+ * Database initialization state
+ */
+enum InitState {
+  INITIALIZING = 'initializing',
+  INITIALIZED = 'initialized',
+  FAILED = 'failed'
+}
+
+/**
+ * DatabaseManager - Singleton class for managing libsql database connections
+ *
+ * Responsibilities:
+ * - Single source of truth for database connection
+ * - Thread-safe initialization with state management
+ * - Automatic migration handling
+ * - Safe connection cleanup
+ * - Error recovery and retry logic
+ * - Windows platform compatibility fixes
+ */
+export class DatabaseManager {
+  private static instance: DatabaseManager | null = null
+
+  private client: Client | null = null
+  private db: LibSQLDatabase<typeof schema> | null = null
+  private state: InitState = InitState.INITIALIZING
+
+  /**
+   * Get the singleton instance (database initialization starts automatically)
+   */
+  public static async getInstance(): Promise<DatabaseManager> {
+    if (DatabaseManager.instance) {
+      return DatabaseManager.instance
+    }
+
+    const instance = new DatabaseManager()
+    await instance.initialize()
+    DatabaseManager.instance = instance
+
+    return instance
+  }
+
+  /**
+   * Perform the actual initialization
+   */
+  public async initialize(): Promise<void> {
+    if (this.state === InitState.INITIALIZED) {
+      return
+    }
+
+    try {
+      logger.info(`Initializing database at: ${dbPath}`)
+
+      // Ensure database directory exists
+      const dbDir = path.dirname(dbPath)
+      if (!fs.existsSync(dbDir)) {
+        logger.info(`Creating database directory: ${dbDir}`)
+        fs.mkdirSync(dbDir, { recursive: true })
+      }
+
+      // Check if database file is corrupted (Windows specific check)
+      if (fs.existsSync(dbPath)) {
+        const stats = fs.statSync(dbPath)
+        if (stats.size === 0) {
+          logger.warn('Database file is empty, removing corrupted file')
+          fs.unlinkSync(dbPath)
+        }
+      }
+
+      // Create client with platform-specific options
+      this.client = createClient({
+        url: `file:${dbPath}`,
+        // intMode: 'number' helps avoid some Windows compatibility issues
+        intMode: 'number'
+      })
+
+      // Create drizzle instance
+      this.db = drizzle(this.client, { schema })
+
+      // Run migrations
+      const migrationService = new MigrationService(this.db, this.client)
+      await migrationService.runMigrations()
+
+      this.state = InitState.INITIALIZED
+      logger.info('Database initialized successfully')
+    } catch (error) {
+      const err = error as Error
+      logger.error('Database initialization failed:', {
+        error: err.message,
+        stack: err.stack
+      })
+
+      // Clean up failed initialization
+      this.cleanupFailedInit()
+
+      // Set failed state
+      this.state = InitState.FAILED
+      throw new Error(`Database initialization failed: ${err.message || 'Unknown error'}`)
+    }
+  }
+
+  /**
+   * Clean up after failed initialization
+   */
+  private cleanupFailedInit(): void {
+    if (this.client) {
+      try {
+        // On Windows, closing a partially initialized client can crash
+        // Wrap in try-catch and ignore errors during cleanup
+        this.client.close()
+      } catch (error) {
+        logger.warn('Failed to close client during cleanup:', error as Error)
+      }
+    }
+    this.client = null
+    this.db = null
+  }
+
+  /**
+   * Get the database instance
+   * Automatically waits for initialization to complete
+   * @throws Error if database initialization failed
+   */
+  public getDatabase(): LibSQLDatabase<typeof schema> {
+    return this.db!
+  }
+
+  /**
+   * Get the raw client (for advanced operations)
+   * Automatically waits for initialization to complete
+   * @throws Error if database initialization failed
+   */
+  public async getClient(): Promise<Client> {
+    return this.client!
+  }
+
+  /**
+   * Check if database is initialized
+   */
+  public isInitialized(): boolean {
+    return this.state === InitState.INITIALIZED
+  }
+}
@@ -7,8 +7,14 @@
 * Schema evolution is handled by Drizzle Kit migrations.
 */

+// Database Manager (Singleton)
+export * from './DatabaseManager'
+
 // Drizzle ORM schemas
 export * from './schema'

 // Repository helpers
 export * from './sessionMessageRepository'
+
+// Migration Service
+export * from './MigrationService'
@@ -15,26 +15,16 @@ import { sessionMessagesTable } from './schema'

 const logger = loggerService.withContext('AgentMessageRepository')

-type TxClient = any
-
 export type PersistUserMessageParams = AgentMessageUserPersistPayload & {
  sessionId: string
  agentSessionId?: string
-  tx?: TxClient
 }

 export type PersistAssistantMessageParams = AgentMessageAssistantPersistPayload & {
  sessionId: string
  agentSessionId: string
-  tx?: TxClient
 }

-type PersistExchangeParams = AgentMessagePersistExchangePayload & {
-  tx?: TxClient
-}
-
-type PersistExchangeResult = AgentMessagePersistExchangeResult
-
 class AgentMessageRepository extends BaseService {
  private static instance: AgentMessageRepository | null = null

@@ -87,17 +77,13 @@ class AgentMessageRepository extends BaseService {
    return deserialized
  }

-  private getWriter(tx?: TxClient): TxClient {
-    return tx ?? this.database
-  }
-
  private async findExistingMessageRow(
-    writer: TxClient,
    sessionId: string,
    role: string,
    messageId: string
  ): Promise<SessionMessageRow | null> {
-    const candidateRows: SessionMessageRow[] = await writer
+    const database = await this.getDatabase()
+    const candidateRows: SessionMessageRow[] = await database
      .select()
      .from(sessionMessagesTable)
      .where(and(eq(sessionMessagesTable.session_id, sessionId), eq(sessionMessagesTable.role, role)))
@@ -122,10 +108,7 @@ class AgentMessageRepository extends BaseService {
  private async upsertMessage(
    params: PersistUserMessageParams | PersistAssistantMessageParams
  ): Promise<AgentSessionMessageEntity> {
-    await AgentMessageRepository.initialize()
-    this.ensureInitialized()
-
-    const { sessionId, agentSessionId = '', payload, metadata, createdAt, tx } = params
+    const { sessionId, agentSessionId = '', payload, metadata, createdAt } = params

    if (!payload?.message?.role) {
      throw new Error('Message payload missing role')
@@ -135,18 +118,18 @@ class AgentMessageRepository extends BaseService {
      throw new Error('Message payload missing id')
    }

-    const writer = this.getWriter(tx)
+    const database = await this.getDatabase()
    const now = createdAt ?? payload.message.createdAt ?? new Date().toISOString()
    const serializedPayload = this.serializeMessage(payload)
    const serializedMetadata = this.serializeMetadata(metadata)

-    const existingRow = await this.findExistingMessageRow(writer, sessionId, payload.message.role, payload.message.id)
+    const existingRow = await this.findExistingMessageRow(sessionId, payload.message.role, payload.message.id)

    if (existingRow) {
      const metadataToPersist = serializedMetadata ?? existingRow.metadata ?? undefined
      const agentSessionToPersist = agentSessionId || existingRow.agent_session_id || ''

-      await writer
+      await database
        .update(sessionMessagesTable)
        .set({
          content: serializedPayload,
@@ -175,7 +158,7 @@ class AgentMessageRepository extends BaseService {
      updated_at: now
    }

-    const [saved] = await writer.insert(sessionMessagesTable).values(insertData).returning()
+    const [saved] = await database.insert(sessionMessagesTable).values(insertData).returning()

    return this.deserialize(saved)
  }
@@ -188,49 +171,38 @@ class AgentMessageRepository extends BaseService {
    return this.upsertMessage(params)
  }

-  async persistExchange(params: PersistExchangeParams): Promise<PersistExchangeResult> {
-    await AgentMessageRepository.initialize()
-    this.ensureInitialized()
-
+  async persistExchange(params: AgentMessagePersistExchangePayload): Promise<AgentMessagePersistExchangeResult> {
    const { sessionId, agentSessionId, user, assistant } = params

-    const result = await this.database.transaction(async (tx) => {
-      const exchangeResult: PersistExchangeResult = {}
+    const exchangeResult: AgentMessagePersistExchangeResult = {}

-      if (user?.payload) {
-        exchangeResult.userMessage = await this.persistUserMessage({
-          sessionId,
-          agentSessionId,
-          payload: user.payload,
-          metadata: user.metadata,
-          createdAt: user.createdAt,
-          tx
-        })
-      }
+    if (user?.payload) {
+      exchangeResult.userMessage = await this.persistUserMessage({
+        sessionId,
+        agentSessionId,
+        payload: user.payload,
+        metadata: user.metadata,
+        createdAt: user.createdAt
+      })
+    }

-      if (assistant?.payload) {
-        exchangeResult.assistantMessage = await this.persistAssistantMessage({
-          sessionId,
-          agentSessionId,
-          payload: assistant.payload,
-          metadata: assistant.metadata,
-          createdAt: assistant.createdAt,
-          tx
-        })
-      }
+    if (assistant?.payload) {
+      exchangeResult.assistantMessage = await this.persistAssistantMessage({
+        sessionId,
+        agentSessionId,
+        payload: assistant.payload,
+        metadata: assistant.metadata,
+        createdAt: assistant.createdAt
+      })
+    }

-      return exchangeResult
-    })
-
-    return result
+    return exchangeResult
  }

  async getSessionHistory(sessionId: string): Promise<AgentPersistedMessage[]> {
-    await AgentMessageRepository.initialize()
-    this.ensureInitialized()
-
    try {
-      const rows = await this.database
+      const database = await this.getDatabase()
+      const rows = await database
        .select()
        .from(sessionMessagesTable)
        .where(eq(sessionMessagesTable.session_id, sessionId))
@@ -32,14 +32,8 @@ export class AgentService extends BaseService {
    return AgentService.instance
  }

-  async initialize(): Promise<void> {
-    await BaseService.initialize()
-  }
-
  // Agent Methods
  async createAgent(req: CreateAgentRequest): Promise<CreateAgentResponse> {
-    this.ensureInitialized()
-
    const id = `agent_${Date.now()}_${Math.random().toString(36).substring(2, 11)}`
    const now = new Date().toISOString()

@@ -75,8 +69,9 @@ export class AgentService extends BaseService {
      updated_at: now
    }

-    await this.database.insert(agentsTable).values(insertData)
-    const result = await this.database.select().from(agentsTable).where(eq(agentsTable.id, id)).limit(1)
+    const database = await this.getDatabase()
+    await database.insert(agentsTable).values(insertData)
+    const result = await database.select().from(agentsTable).where(eq(agentsTable.id, id)).limit(1)
    if (!result[0]) {
      throw new Error('Failed to create agent')
    }
@@ -86,9 +81,8 @@ export class AgentService extends BaseService {
  }

  async getAgent(id: string): Promise<GetAgentResponse | null> {
-    this.ensureInitialized()
-
-    const result = await this.database.select().from(agentsTable).where(eq(agentsTable.id, id)).limit(1)
+    const database = await this.getDatabase()
+    const result = await database.select().from(agentsTable).where(eq(agentsTable.id, id)).limit(1)

    if (!result[0]) {
      return null
@@ -118,9 +112,9 @@ export class AgentService extends BaseService {
  }

  async listAgents(options: ListOptions = {}): Promise<{ agents: AgentEntity[]; total: number }> {
-    this.ensureInitialized() // Build query with pagination
-
-    const totalResult = await this.database.select({ count: count() }).from(agentsTable)
+    // Build query with pagination
+    const database = await this.getDatabase()
+    const totalResult = await database.select({ count: count() }).from(agentsTable)

    const sortBy = options.sortBy || 'created_at'
    const orderBy = options.orderBy || 'desc'
@@ -128,7 +122,7 @@ export class AgentService extends BaseService {
    const sortField = agentsTable[sortBy]
    const orderFn = orderBy === 'asc' ? asc : desc

-    const baseQuery = this.database.select().from(agentsTable).orderBy(orderFn(sortField))
+    const baseQuery = database.select().from(agentsTable).orderBy(orderFn(sortField))

    const result =
      options.limit !== undefined
@@ -151,8 +145,6 @@ export class AgentService extends BaseService {
    updates: UpdateAgentRequest,
    options: { replace?: boolean } = {}
  ): Promise<UpdateAgentResponse | null> {
-    this.ensureInitialized()
-
    // Check if agent exists
    const existing = await this.getAgent(id)
    if (!existing) {
@@ -195,22 +187,21 @@ export class AgentService extends BaseService {
      }
    }

-    await this.database.update(agentsTable).set(updateData).where(eq(agentsTable.id, id))
+    const database = await this.getDatabase()
+    await database.update(agentsTable).set(updateData).where(eq(agentsTable.id, id))
    return await this.getAgent(id)
  }

  async deleteAgent(id: string): Promise<boolean> {
-    this.ensureInitialized()
-
-    const result = await this.database.delete(agentsTable).where(eq(agentsTable.id, id))
+    const database = await this.getDatabase()
+    const result = await database.delete(agentsTable).where(eq(agentsTable.id, id))

    return result.rowsAffected > 0
  }

  async agentExists(id: string): Promise<boolean> {
-    this.ensureInitialized()
-
-    const result = await this.database
+    const database = await this.getDatabase()
+    const result = await database
      .select({ id: agentsTable.id })
      .from(agentsTable)
      .where(eq(agentsTable.id, id))
@@ -104,14 +104,9 @@ export class SessionMessageService extends BaseService {
    return SessionMessageService.instance
  }

-  async initialize(): Promise<void> {
-    await BaseService.initialize()
-  }
-
  async sessionMessageExists(id: number): Promise<boolean> {
-    this.ensureInitialized()
-
-    const result = await this.database
+    const database = await this.getDatabase()
+    const result = await database
      .select({ id: sessionMessagesTable.id })
      .from(sessionMessagesTable)
      .where(eq(sessionMessagesTable.id, id))
@@ -124,10 +119,9 @@ export class SessionMessageService extends BaseService {
    sessionId: string,
    options: ListOptions = {}
  ): Promise<{ messages: AgentSessionMessageEntity[] }> {
-    this.ensureInitialized()
-
    // Get messages with pagination
-    const baseQuery = this.database
+    const database = await this.getDatabase()
+    const baseQuery = database
      .select()
      .from(sessionMessagesTable)
      .where(eq(sessionMessagesTable.session_id, sessionId))
@@ -146,9 +140,8 @@ export class SessionMessageService extends BaseService {
  }

  async deleteSessionMessage(sessionId: string, messageId: number): Promise<boolean> {
-    this.ensureInitialized()
-
-    const result = await this.database
+    const database = await this.getDatabase()
+    const result = await database
      .delete(sessionMessagesTable)
      .where(and(eq(sessionMessagesTable.id, messageId), eq(sessionMessagesTable.session_id, sessionId)))

@@ -160,8 +153,6 @@ export class SessionMessageService extends BaseService {
    messageData: CreateSessionMessageRequest,
    abortController: AbortController
  ): Promise<SessionStreamResult> {
-    this.ensureInitialized()
-
    return await this.startSessionMessageStream(session, messageData, abortController)
  }

@@ -270,10 +261,9 @@ export class SessionMessageService extends BaseService {
  }

  private async getLastAgentSessionId(sessionId: string): Promise<string> {
-    this.ensureInitialized()
-
    try {
-      const result = await this.database
+      const database = await this.getDatabase()
+      const result = await database
        .select({ agent_session_id: sessionMessagesTable.agent_session_id })
        .from(sessionMessagesTable)
        .where(and(eq(sessionMessagesTable.session_id, sessionId), not(eq(sessionMessagesTable.agent_session_id, ''))))
@@ -30,10 +30,6 @@ export class SessionService extends BaseService {
    return SessionService.instance
  }

-  async initialize(): Promise<void> {
-    await BaseService.initialize()
-  }
-
  /**
   * Override BaseService.listSlashCommands to merge builtin and plugin commands
   */
@@ -84,13 +80,12 @@ export class SessionService extends BaseService {
    agentId: string,
    req: Partial<CreateSessionRequest> = {}
  ): Promise<GetAgentSessionResponse | null> {
-    this.ensureInitialized()
-
    // Validate agent exists - we'll need to import AgentService for this check
    // For now, we'll skip this validation to avoid circular dependencies
    // The database foreign key constraint will handle this

-    const agents = await this.database.select().from(agentsTable).where(eq(agentsTable.id, agentId)).limit(1)
+    const database = await this.getDatabase()
+    const agents = await database.select().from(agentsTable).where(eq(agentsTable.id, agentId)).limit(1)
    if (!agents[0]) {
      throw new Error('Agent not found')
    }
@@ -135,9 +130,10 @@ export class SessionService extends BaseService {
      updated_at: now
    }

-    await this.database.insert(sessionsTable).values(insertData)
+    const db = await this.getDatabase()
+    await db.insert(sessionsTable).values(insertData)

-    const result = await this.database.select().from(sessionsTable).where(eq(sessionsTable.id, id)).limit(1)
+    const result = await db.select().from(sessionsTable).where(eq(sessionsTable.id, id)).limit(1)

    if (!result[0]) {
      throw new Error('Failed to create session')
@@ -148,9 +144,8 @@ export class SessionService extends BaseService {
  }

  async getSession(agentId: string, id: string): Promise<GetAgentSessionResponse | null> {
-    this.ensureInitialized()
-
-    const result = await this.database
+    const database = await this.getDatabase()
+    const result = await database
      .select()
      .from(sessionsTable)
      .where(and(eq(sessionsTable.id, id), eq(sessionsTable.agent_id, agentId)))
@@ -176,8 +171,6 @@ export class SessionService extends BaseService {
    agentId?: string,
    options: ListOptions = {}
  ): Promise<{ sessions: AgentSessionEntity[]; total: number }> {
-    this.ensureInitialized()
-
    // Build where conditions
    const whereConditions: SQL[] = []
    if (agentId) {
@@ -192,16 +185,13 @@ export class SessionService extends BaseService {
          : undefined

    // Get total count
-    const totalResult = await this.database.select({ count: count() }).from(sessionsTable).where(whereClause)
+    const database = await this.getDatabase()
+    const totalResult = await database.select({ count: count() }).from(sessionsTable).where(whereClause)

    const total = totalResult[0].count

    // Build list query with pagination - sort by updated_at descending (latest first)
-    const baseQuery = this.database
-      .select()
-      .from(sessionsTable)
-      .where(whereClause)
-      .orderBy(desc(sessionsTable.updated_at))
+    const baseQuery = database.select().from(sessionsTable).where(whereClause).orderBy(desc(sessionsTable.updated_at))

    const result =
      options.limit !== undefined
@@ -220,8 +210,6 @@ export class SessionService extends BaseService {
    id: string,
    updates: UpdateSessionRequest
  ): Promise<UpdateSessionResponse | null> {
-    this.ensureInitialized()
-
    // Check if session exists
    const existing = await this.getSession(agentId, id)
    if (!existing) {
@@ -262,15 +250,15 @@ export class SessionService extends BaseService {
      }
    }

-    await this.database.update(sessionsTable).set(updateData).where(eq(sessionsTable.id, id))
+    const database = await this.getDatabase()
+    await database.update(sessionsTable).set(updateData).where(eq(sessionsTable.id, id))

    return await this.getSession(agentId, id)
  }

  async deleteSession(agentId: string, id: string): Promise<boolean> {
-    this.ensureInitialized()
-
-    const result = await this.database
+    const database = await this.getDatabase()
+    const result = await database
      .delete(sessionsTable)
      .where(and(eq(sessionsTable.id, id), eq(sessionsTable.agent_id, agentId)))

@@ -278,9 +266,8 @@ export class SessionService extends BaseService {
  }

  async sessionExists(agentId: string, id: string): Promise<boolean> {
-    this.ensureInitialized()
-
-    const result = await this.database
+    const database = await this.getDatabase()
+    const result = await database
      .select({ id: sessionsTable.id })
      .from(sessionsTable)
      .where(and(eq(sessionsTable.id, id), eq(sessionsTable.agent_id, agentId)))
@@ -21,11 +21,16 @@ describe('stripLocalCommandTags', () => {
      '<local-command-stdout>line1</local-command-stdout>\nkeep\n<local-command-stderr>Error</local-command-stderr>'
    expect(stripLocalCommandTags(input)).toBe('line1\nkeep\nError')
  })
+
+  it('if no tags present, returns original string', () => {
+    const input = 'just some normal text'
+    expect(stripLocalCommandTags(input)).toBe(input)
+  })
 })

 describe('Claude → AiSDK transform', () => {
  it('handles tool call streaming lifecycle', () => {
-    const state = new ClaudeStreamState()
+    const state = new ClaudeStreamState({ agentSessionId: baseStreamMetadata.session_id })
    const parts: ReturnType<typeof transformSDKMessageToStreamParts>[number][] = []

    const messages: SDKMessage[] = [
@@ -182,14 +187,119 @@ describe('Claude → AiSDK transform', () => {
      (typeof parts)[number],
      { type: 'tool-result' }
    >
-    expect(toolResult.toolCallId).toBe('tool-1')
+    expect(toolResult.toolCallId).toBe('session-123:tool-1')
    expect(toolResult.toolName).toBe('Bash')
    expect(toolResult.input).toEqual({ command: 'ls' })
    expect(toolResult.output).toBe('ok')
  })

+  it('handles tool calls without streaming events (no content_block_start/stop)', () => {
+    const state = new ClaudeStreamState({ agentSessionId: '12344' })
+    const parts: ReturnType<typeof transformSDKMessageToStreamParts>[number][] = []
+
+    const messages: SDKMessage[] = [
+      {
+        ...baseStreamMetadata,
+        type: 'assistant',
+        uuid: uuid(20),
+        message: {
+          id: 'msg-tool-no-stream',
+          type: 'message',
+          role: 'assistant',
+          model: 'claude-test',
+          content: [
+            {
+              type: 'tool_use',
+              id: 'tool-read',
+              name: 'Read',
+              input: { file_path: '/test.txt' }
+            },
+            {
+              type: 'tool_use',
+              id: 'tool-bash',
+              name: 'Bash',
+              input: { command: 'ls -la' }
+            }
+          ],
+          stop_reason: 'tool_use',
+          stop_sequence: null,
+          usage: {
+            input_tokens: 10,
+            output_tokens: 20
+          }
+        }
+      } as unknown as SDKMessage,
+      {
+        ...baseStreamMetadata,
+        type: 'user',
+        uuid: uuid(21),
+        message: {
+          role: 'user',
+          content: [
+            {
+              type: 'tool_result',
+              tool_use_id: 'tool-read',
+              content: 'file contents',
+              is_error: false
+            }
+          ]
+        }
+      } as SDKMessage,
+      {
+        ...baseStreamMetadata,
+        type: 'user',
+        uuid: uuid(22),
+        message: {
+          role: 'user',
+          content: [
+            {
+              type: 'tool_result',
+              tool_use_id: 'tool-bash',
+              content: 'total 42\n...',
+              is_error: false
+            }
+          ]
+        }
+      } as SDKMessage
+    ]
+
+    for (const message of messages) {
+      const transformed = transformSDKMessageToStreamParts(message, state)
+      parts.push(...transformed)
+    }
+
+    const types = parts.map((part) => part.type)
+    expect(types).toEqual(['tool-call', 'tool-call', 'tool-result', 'tool-result'])
+
+    const toolCalls = parts.filter((part) => part.type === 'tool-call') as Extract<
+      (typeof parts)[number],
+      { type: 'tool-call' }
+    >[]
+    expect(toolCalls).toHaveLength(2)
+    expect(toolCalls[0].toolName).toBe('Read')
+    expect(toolCalls[0].toolCallId).toBe('12344:tool-read')
+    expect(toolCalls[1].toolName).toBe('Bash')
+    expect(toolCalls[1].toolCallId).toBe('12344:tool-bash')
+
+    const toolResults = parts.filter((part) => part.type === 'tool-result') as Extract<
+      (typeof parts)[number],
+      { type: 'tool-result' }
+    >[]
+    expect(toolResults).toHaveLength(2)
+    // This is the key assertion - toolName should NOT be 'unknown'
+    expect(toolResults[0].toolName).toBe('Read')
+    expect(toolResults[0].toolCallId).toBe('12344:tool-read')
+    expect(toolResults[0].input).toEqual({ file_path: '/test.txt' })
+    expect(toolResults[0].output).toBe('file contents')
+
+    expect(toolResults[1].toolName).toBe('Bash')
+    expect(toolResults[1].toolCallId).toBe('12344:tool-bash')
+    expect(toolResults[1].input).toEqual({ command: 'ls -la' })
+    expect(toolResults[1].output).toBe('total 42\n...')
+  })
+
  it('handles streaming text completion', () => {
-    const state = new ClaudeStreamState()
+    const state = new ClaudeStreamState({ agentSessionId: baseStreamMetadata.session_id })
    const parts: ReturnType<typeof transformSDKMessageToStreamParts>[number][] = []

    const messages: SDKMessage[] = [
@@ -300,4 +410,87 @@ describe('Claude → AiSDK transform', () => {
    expect(finishStep.finishReason).toBe('stop')
    expect(finishStep.usage).toEqual({ inputTokens: 2, outputTokens: 4, totalTokens: 6 })
  })
+
+  it('emits fallback text when Claude sends a snapshot instead of deltas', () => {
+    const state = new ClaudeStreamState({ agentSessionId: '12344' })
+    const parts: ReturnType<typeof transformSDKMessageToStreamParts>[number][] = []
+
+    const messages: SDKMessage[] = [
+      {
+        ...baseStreamMetadata,
+        type: 'stream_event',
+        uuid: uuid(30),
+        event: {
+          type: 'message_start',
+          message: {
+            id: 'msg-fallback',
+            type: 'message',
+            role: 'assistant',
+            model: 'claude-test',
+            content: [],
+            stop_reason: null,
+            stop_sequence: null,
+            usage: {}
+          }
+        }
+      } as unknown as SDKMessage,
+      {
+        ...baseStreamMetadata,
+        type: 'stream_event',
+        uuid: uuid(31),
+        event: {
+          type: 'content_block_start',
+          index: 0,
+          content_block: {
+            type: 'text',
+            text: ''
+          }
+        }
+      } as unknown as SDKMessage,
+      {
+        ...baseStreamMetadata,
+        type: 'assistant',
+        uuid: uuid(32),
+        message: {
+          id: 'msg-fallback-content',
+          type: 'message',
+          role: 'assistant',
+          model: 'claude-test',
+          content: [
+            {
+              type: 'text',
+              text: 'Final answer without streaming deltas.'
+            }
+          ],
+          stop_reason: 'end_turn',
+          stop_sequence: null,
+          usage: {
+            input_tokens: 3,
+            output_tokens: 7
+          }
+        }
+      } as unknown as SDKMessage
+    ]
+
+    for (const message of messages) {
+      const transformed = transformSDKMessageToStreamParts(message, state)
+      parts.push(...transformed)
+    }
+
+    const types = parts.map((part) => part.type)
+    expect(types).toEqual(['start-step', 'text-start', 'text-delta', 'text-end', 'finish-step'])
+
+    const delta = parts.find((part) => part.type === 'text-delta') as Extract<
+      (typeof parts)[number],
+      { type: 'text-delta' }
+    >
+    expect(delta.text).toBe('Final answer without streaming deltas.')
+
+    const finish = parts.find((part) => part.type === 'finish-step') as Extract<
+      (typeof parts)[number],
+      { type: 'finish-step' }
+    >
+    expect(finish.usage).toEqual({ inputTokens: 3, outputTokens: 7, totalTokens: 10 })
+    expect(finish.finishReason).toBe('stop')
+  })
 })
@@ -10,8 +10,21 @@
 * Every Claude turn gets its own instance. `resetStep` should be invoked once the finish event has
 * been emitted to avoid leaking state into the next turn.
 */
+import { loggerService } from '@logger'
 import type { FinishReason, LanguageModelUsage, ProviderMetadata } from 'ai'

+/**
+ * Builds a namespaced tool call ID by combining session ID with raw tool call ID.
+ * This ensures tool calls from different sessions don't conflict even if they have
+ * the same raw ID from the SDK.
+ *
+ * @param sessionId - The agent session ID
+ * @param rawToolCallId - The raw tool call ID from SDK (e.g., "WebFetch_0")
+ */
+export function buildNamespacedToolCallId(sessionId: string, rawToolCallId: string): string {
+  return `${sessionId}:${rawToolCallId}`
+}
+
 /**
 * Shared fields for every block that Claude can stream (text, reasoning, tool).
 */
@@ -34,6 +47,7 @@ type ReasoningBlockState = BaseBlockState & {
 type ToolBlockState = BaseBlockState & {
  kind: 'tool'
  toolCallId: string
+  rawToolCallId: string
  toolName: string
  inputBuffer: string
  providerMetadata?: ProviderMetadata
@@ -48,12 +62,17 @@ type PendingUsageState = {
 }

 type PendingToolCall = {
+  rawToolCallId: string
  toolCallId: string
  toolName: string
  input: unknown
  providerMetadata?: ProviderMetadata
 }

+type ClaudeStreamStateOptions = {
+  agentSessionId: string
+}
+
 /**
 * Tracks the lifecycle of Claude streaming blocks (text, thinking, tool calls)
 * across individual websocket events. The transformer relies on this class to
@@ -61,12 +80,20 @@ type PendingToolCall = {
 * usage/finish metadata once Anthropic closes a message.
 */
 export class ClaudeStreamState {
+  private logger
+  private readonly agentSessionId: string
  private blocksByIndex = new Map<number, BlockState>()
-  private toolIndexById = new Map<string, number>()
+  private toolIndexByNamespacedId = new Map<string, number>()
  private pendingUsage: PendingUsageState = {}
  private pendingToolCalls = new Map<string, PendingToolCall>()
  private stepActive = false

+  constructor(options: ClaudeStreamStateOptions) {
+    this.logger = loggerService.withContext('ClaudeStreamState')
+    this.agentSessionId = options.agentSessionId
+    this.logger.silly('ClaudeStreamState', options)
+  }
+
  /** Marks the beginning of a new AiSDK step. */
  beginStep(): void {
    this.stepActive = true
@@ -104,19 +131,21 @@ export class ClaudeStreamState {
  /** Caches tool metadata so subsequent input deltas and results can find it. */
  openToolBlock(
    index: number,
-    params: { toolCallId: string; toolName: string; providerMetadata?: ProviderMetadata }
+    params: { rawToolCallId: string; toolName: string; providerMetadata?: ProviderMetadata }
  ): ToolBlockState {
+    const toolCallId = buildNamespacedToolCallId(this.agentSessionId, params.rawToolCallId)
    const block: ToolBlockState = {
      kind: 'tool',
-      id: params.toolCallId,
+      id: toolCallId,
      index,
-      toolCallId: params.toolCallId,
+      toolCallId,
+      rawToolCallId: params.rawToolCallId,
      toolName: params.toolName,
      inputBuffer: '',
      providerMetadata: params.providerMetadata
    }
    this.blocksByIndex.set(index, block)
-    this.toolIndexById.set(params.toolCallId, index)
+    this.toolIndexByNamespacedId.set(toolCallId, index)
    return block
  }

@@ -124,14 +153,32 @@ export class ClaudeStreamState {
    return this.blocksByIndex.get(index)
  }

+  getFirstOpenTextBlock(): TextBlockState | undefined {
+    const candidates: TextBlockState[] = []
+    for (const block of this.blocksByIndex.values()) {
+      if (block.kind === 'text') {
+        candidates.push(block)
+      }
+    }
+    if (candidates.length === 0) {
+      return undefined
+    }
+    candidates.sort((a, b) => a.index - b.index)
+    return candidates[0]
+  }
+
  getToolBlockById(toolCallId: string): ToolBlockState | undefined {
-    const index = this.toolIndexById.get(toolCallId)
+    const index = this.toolIndexByNamespacedId.get(toolCallId)
    if (index === undefined) return undefined
    const block = this.blocksByIndex.get(index)
    if (!block || block.kind !== 'tool') return undefined
    return block
  }

+  getToolBlockByRawId(rawToolCallId: string): ToolBlockState | undefined {
+    return this.getToolBlockById(buildNamespacedToolCallId(this.agentSessionId, rawToolCallId))
+  }
+
  /** Appends streamed text to a text block, returning the updated state when present. */
  appendTextDelta(index: number, text: string): TextBlockState | undefined {
    const block = this.blocksByIndex.get(index)
@@ -158,10 +205,12 @@ export class ClaudeStreamState {

  /** Records a tool call to be consumed once its result arrives from the user. */
  registerToolCall(
-    toolCallId: string,
+    rawToolCallId: string,
    payload: { toolName: string; input: unknown; providerMetadata?: ProviderMetadata }
  ): void {
-    this.pendingToolCalls.set(toolCallId, {
+    const toolCallId = buildNamespacedToolCallId(this.agentSessionId, rawToolCallId)
+    this.pendingToolCalls.set(rawToolCallId, {
+      rawToolCallId,
      toolCallId,
      toolName: payload.toolName,
      input: payload.input,
@@ -170,10 +219,10 @@ export class ClaudeStreamState {
  }

  /** Retrieves and clears the buffered tool call metadata for the given id. */
-  consumePendingToolCall(toolCallId: string): PendingToolCall | undefined {
-    const entry = this.pendingToolCalls.get(toolCallId)
+  consumePendingToolCall(rawToolCallId: string): PendingToolCall | undefined {
+    const entry = this.pendingToolCalls.get(rawToolCallId)
    if (entry) {
-      this.pendingToolCalls.delete(toolCallId)
+      this.pendingToolCalls.delete(rawToolCallId)
    }
    return entry
  }
@@ -182,13 +231,13 @@ export class ClaudeStreamState {
   * Persists the final input payload for a tool block once the provider signals
   * completion so that downstream tool results can reference the original call.
   */
-  completeToolBlock(toolCallId: string, input: unknown, providerMetadata?: ProviderMetadata): void {
+  completeToolBlock(toolCallId: string, toolName: string, input: unknown, providerMetadata?: ProviderMetadata): void {
+    const block = this.getToolBlockByRawId(toolCallId)
    this.registerToolCall(toolCallId, {
-      toolName: this.getToolBlockById(toolCallId)?.toolName ?? 'unknown',
+      toolName,
      input,
      providerMetadata
    })
-    const block = this.getToolBlockById(toolCallId)
    if (block) {
      block.resolvedInput = input
    }
@@ -200,7 +249,7 @@ export class ClaudeStreamState {
    if (!block) return undefined
    this.blocksByIndex.delete(index)
    if (block.kind === 'tool') {
-      this.toolIndexById.delete(block.toolCallId)
+      this.toolIndexByNamespacedId.delete(block.toolCallId)
    }
    return block
  }
@@ -227,7 +276,7 @@ export class ClaudeStreamState {
  /** Drops cached block metadata for the currently active message. */
  resetBlocks(): void {
    this.blocksByIndex.clear()
-    this.toolIndexById.clear()
+    this.toolIndexByNamespacedId.clear()
  }

  /** Resets the entire step lifecycle after emitting a terminal frame. */
@@ -236,6 +285,10 @@ export class ClaudeStreamState {
    this.resetPendingUsage()
    this.stepActive = false
  }
+
+  getNamespacedToolCallId(rawToolCallId: string): string {
+    return buildNamespacedToolCallId(this.agentSessionId, rawToolCallId)
+  }
 }

 export type { PendingToolCall }
@@ -2,7 +2,14 @@
 import { EventEmitter } from 'node:events'
 import { createRequire } from 'node:module'

-import type { CanUseTool, McpHttpServerConfig, Options, SDKMessage } from '@anthropic-ai/claude-agent-sdk'
+import type {
+  CanUseTool,
+  HookCallback,
+  McpHttpServerConfig,
+  Options,
+  PreToolUseHookInput,
+  SDKMessage
+} from '@anthropic-ai/claude-agent-sdk'
 import { query } from '@anthropic-ai/claude-agent-sdk'
 import { loggerService } from '@logger'
 import { config as apiConfigService } from '@main/apiServer/config'
@@ -13,6 +20,7 @@ import { app } from 'electron'
 import type { GetAgentSessionResponse } from '../..'
 import type { AgentServiceInterface, AgentStream, AgentStreamEvent } from '../../interfaces/AgentStreamInterface'
 import { sessionService } from '../SessionService'
+import { buildNamespacedToolCallId } from './claude-stream-state'
 import { promptForToolApproval } from './tool-permissions'
 import { ClaudeStreamState, transformSDKMessageToStreamParts } from './transform'

@@ -150,7 +158,67 @@ class ClaudeCodeService implements AgentServiceInterface {
        return { behavior: 'allow', updatedInput: input }
      }

-      return promptForToolApproval(toolName, input, options)
+      return promptForToolApproval(toolName, input, {
+        ...options,
+        toolCallId: buildNamespacedToolCallId(session.id, options.toolUseID)
+      })
+    }
+
+    const preToolUseHook: HookCallback = async (input, toolUseID, options) => {
+      // Type guard to ensure we're handling PreToolUse event
+      if (input.hook_event_name !== 'PreToolUse') {
+        return {}
+      }
+
+      const hookInput = input as PreToolUseHookInput
+      const toolName = hookInput.tool_name
+
+      logger.debug('PreToolUse hook triggered', {
+        session_id: hookInput.session_id,
+        tool_name: hookInput.tool_name,
+        tool_use_id: toolUseID,
+        tool_input: hookInput.tool_input,
+        cwd: hookInput.cwd,
+        permission_mode: hookInput.permission_mode,
+        autoAllowTools: autoAllowTools
+      })
+
+      if (options?.signal?.aborted) {
+        logger.debug('PreToolUse hook signal already aborted; skipping tool use', {
+          tool_name: hookInput.tool_name
+        })
+        return {}
+      }
+
+      // handle auto approved tools since it never triggers canUseTool
+      const normalizedToolName = normalizeToolName(toolName)
+      if (toolUseID) {
+        const bypassAll = input.permission_mode === 'bypassPermissions'
+        const autoAllowed = autoAllowTools.has(toolName) || autoAllowTools.has(normalizedToolName)
+        if (bypassAll || autoAllowed) {
+          const namespacedToolCallId = buildNamespacedToolCallId(session.id, toolUseID)
+          logger.debug('handling auto approved tools', {
+            toolName,
+            normalizedToolName,
+            namespacedToolCallId,
+            permission_mode: input.permission_mode,
+            autoAllowTools
+          })
+          const isRecord = (v: unknown): v is Record<string, unknown> => {
+            return !!v && typeof v === 'object' && !Array.isArray(v)
+          }
+          const toolInput = isRecord(input.tool_input) ? input.tool_input : {}
+
+          await promptForToolApproval(toolName, toolInput, {
+            ...options,
+            toolCallId: namespacedToolCallId,
+            autoApprove: true
+          })
+        }
+      }
+
+      // Return to proceed without modification
+      return {}
    }

    // Build SDK options from parameters
@@ -176,7 +244,14 @@ class ClaudeCodeService implements AgentServiceInterface {
      permissionMode: session.configuration?.permission_mode,
      maxTurns: session.configuration?.max_turns,
      allowedTools: session.allowed_tools,
-      canUseTool
+      canUseTool,
+      hooks: {
+        PreToolUse: [
+          {
+            hooks: [preToolUseHook]
+          }
+        ]
+      }
    }

    if (session.accessible_paths.length > 1) {
@@ -346,7 +421,7 @@ class ClaudeCodeService implements AgentServiceInterface {
    const jsonOutput: SDKMessage[] = []
    let hasCompleted = false
    const startTime = Date.now()
-    const streamState = new ClaudeStreamState()
+    const streamState = new ClaudeStreamState({ agentSessionId: sessionId })

    try {
      for await (const message of query({ prompt: promptStream, options })) {
@@ -410,23 +485,6 @@ class ClaudeCodeService implements AgentServiceInterface {
          }
        }

-        if (message.type === 'assistant' || message.type === 'user') {
-          logger.silly('claude response', {
-            message,
-            content: JSON.stringify(message.message.content)
-          })
-        } else if (message.type === 'stream_event') {
-          // logger.silly('Claude stream event', {
-          //   message,
-          //   event: JSON.stringify(message.event)
-          // })
-        } else {
-          logger.silly('Claude response', {
-            message,
-            event: JSON.stringify(message)
-          })
-        }
-
        const chunks = transformSDKMessageToStreamParts(message, streamState)
        for (const chunk of chunks) {
          stream.emit('data', {
@@ -31,12 +31,14 @@ type PendingPermissionRequest = {
  abortListener?: () => void
  originalInput: Record<string, unknown>
  toolName: string
+  toolCallId?: string
 }

 type RendererPermissionRequestPayload = {
  requestId: string
  toolName: string
  toolId: string
+  toolCallId: string
  description?: string
  requiresPermissions: boolean
  input: Record<string, unknown>
@@ -44,6 +46,7 @@ type RendererPermissionRequestPayload = {
  createdAt: number
  expiresAt: number
  suggestions: PermissionUpdate[]
+  autoApprove?: boolean
 }

 type RendererPermissionResultPayload = {
@@ -51,6 +54,7 @@ type RendererPermissionResultPayload = {
  behavior: ToolPermissionBehavior
  message?: string
  reason: 'response' | 'timeout' | 'aborted' | 'no-window'
+  toolCallId?: string
 }

 const pendingRequests = new Map<string, PendingPermissionRequest>()
@@ -144,7 +148,8 @@ const finalizeRequest = (
    requestId,
    behavior: update.behavior,
    message: update.behavior === 'deny' ? update.message : undefined,
-    reason
+    reason,
+    toolCallId: pending.toolCallId
  }

  const dispatched = broadcastToRenderer(IpcChannel.AgentToolPermission_Result, resultPayload)
@@ -206,10 +211,20 @@ const ensureIpcHandlersRegistered = () => {
  })
 }

+type PromptForToolApprovalOptions = {
+  signal: AbortSignal
+  suggestions?: PermissionUpdate[]
+  autoApprove?: boolean
+
+  // NOTICE: This ID is namespaced with session ID, not the raw SDK tool call ID.
+  // Format: `${sessionId}:${rawToolCallId}`, e.g., `session_123:WebFetch_0`
+  toolCallId: string
+}
+
 export async function promptForToolApproval(
  toolName: string,
  input: Record<string, unknown>,
-  options?: { signal: AbortSignal; suggestions?: PermissionUpdate[] }
+  options: PromptForToolApprovalOptions
 ): Promise<PermissionResult> {
  if (shouldAutoApproveTools) {
    logger.debug('promptForToolApproval auto-approving tool for test', {
@@ -245,6 +260,7 @@ export async function promptForToolApproval(
  logger.info('Requesting user approval for tool usage', {
    requestId,
    toolName,
+    toolCallId: options.toolCallId,
    description: toolMetadata?.description
  })

@@ -252,13 +268,15 @@ export async function promptForToolApproval(
    requestId,
    toolName,
    toolId: toolMetadata?.id ?? toolName,
+    toolCallId: options.toolCallId,
    description: toolMetadata?.description,
    requiresPermissions: toolMetadata?.requirePermissions ?? false,
    input: sanitizedInput,
    inputPreview,
    createdAt,
    expiresAt,
-    suggestions: sanitizedSuggestions
+    suggestions: sanitizedSuggestions,
+    autoApprove: options.autoApprove
  }

  const defaultDenyUpdate: PermissionResult = { behavior: 'deny', message: 'Tool request aborted before user decision' }
@@ -266,6 +284,7 @@ export async function promptForToolApproval(
  logger.debug('Registering tool permission request', {
    requestId,
    toolName,
+    toolCallId: options.toolCallId,
    requiresPermissions: requestPayload.requiresPermissions,
    timeoutMs: TOOL_APPROVAL_TIMEOUT_MS,
    suggestionCount: sanitizedSuggestions.length
@@ -273,7 +292,11 @@ export async function promptForToolApproval(

  return new Promise<PermissionResult>((resolve) => {
    const timeout = setTimeout(() => {
-      logger.info('User tool permission request timed out', { requestId, toolName })
+      logger.info('User tool permission request timed out', {
+        requestId,
+        toolName,
+        toolCallId: options.toolCallId
+      })
      finalizeRequest(requestId, { behavior: 'deny', message: 'Timed out waiting for approval' }, 'timeout')
    }, TOOL_APPROVAL_TIMEOUT_MS)

@@ -282,12 +305,17 @@ export async function promptForToolApproval(
      timeout,
      originalInput: sanitizedInput,
      toolName,
-      signal: options?.signal
+      signal: options?.signal,
+      toolCallId: options.toolCallId
    }

    if (options?.signal) {
      const abortListener = () => {
-        logger.info('Tool permission request aborted before user responded', { requestId, toolName })
+        logger.info('Tool permission request aborted before user responded', {
+          requestId,
+          toolName,
+          toolCallId: options.toolCallId
+        })
        finalizeRequest(requestId, defaultDenyUpdate, 'aborted')
      }

@@ -110,7 +110,7 @@ const sdkMessageToProviderMetadata = (message: SDKMessage): ProviderMetadata =>
 * blocks across calls so that incremental deltas can be correlated correctly.
 */
 export function transformSDKMessageToStreamParts(sdkMessage: SDKMessage, state: ClaudeStreamState): AgentStreamPart[] {
-  logger.silly('Transforming SDKMessage', { message: sdkMessage })
+  logger.silly('Transforming SDKMessage', { message: JSON.stringify(sdkMessage) })
  switch (sdkMessage.type) {
    case 'assistant':
      return handleAssistantMessage(sdkMessage, state)
@@ -186,14 +186,13 @@ function handleAssistantMessage(

  for (const block of content) {
    switch (block.type) {
-      case 'text':
-        if (!isStreamingActive) {
-          const sanitizedText = stripLocalCommandTags(block.text)
-          if (sanitizedText) {
-            textBlocks.push(sanitizedText)
-          }
+      case 'text': {
+        const sanitizedText = stripLocalCommandTags(block.text)
+        if (sanitizedText) {
+          textBlocks.push(sanitizedText)
        }
        break
+      }
      case 'tool_use':
        handleAssistantToolUse(block as ToolUseContent, providerMetadata, state, chunks)
        break
@@ -203,7 +202,16 @@ function handleAssistantMessage(
    }
  }

-  if (!isStreamingActive && textBlocks.length > 0) {
+  if (textBlocks.length === 0) {
+    return chunks
+  }
+
+  const combinedText = textBlocks.join('')
+  if (!combinedText) {
+    return chunks
+  }
+
+  if (!isStreamingActive) {
    const id = message.uuid?.toString() || generateMessageId()
    state.beginStep()
    chunks.push({
@@ -219,7 +227,7 @@ function handleAssistantMessage(
    chunks.push({
      type: 'text-delta',
      id,
-      text: textBlocks.join(''),
+      text: combinedText,
      providerMetadata
    })
    chunks.push({
@@ -230,7 +238,27 @@ function handleAssistantMessage(
    return finalizeNonStreamingStep(message, state, chunks)
  }

-  return chunks
+  const existingTextBlock = state.getFirstOpenTextBlock()
+  const fallbackId = existingTextBlock?.id || message.uuid?.toString() || generateMessageId()
+  if (!existingTextBlock) {
+    chunks.push({
+      type: 'text-start',
+      id: fallbackId,
+      providerMetadata
+    })
+  }
+  chunks.push({
+    type: 'text-delta',
+    id: fallbackId,
+    text: combinedText,
+    providerMetadata
+  })
+  chunks.push({
+    type: 'text-end',
+    id: fallbackId,
+    providerMetadata
+  })
+  return finalizeNonStreamingStep(message, state, chunks)
 }

 /**
@@ -243,15 +271,16 @@ function handleAssistantToolUse(
  state: ClaudeStreamState,
  chunks: AgentStreamPart[]
 ): void {
+  const toolCallId = state.getNamespacedToolCallId(block.id)
  chunks.push({
    type: 'tool-call',
-    toolCallId: block.id,
+    toolCallId,
    toolName: block.name,
    input: block.input,
    providerExecuted: true,
    providerMetadata
  })
-  state.completeToolBlock(block.id, block.input, providerMetadata)
+  state.completeToolBlock(block.id, block.name, block.input, providerMetadata)
 }

 /**
@@ -331,10 +360,11 @@ function handleUserMessage(
    if (block.type === 'tool_result') {
      const toolResult = block as ToolResultContent
      const pendingCall = state.consumePendingToolCall(toolResult.tool_use_id)
+      const toolCallId = pendingCall?.toolCallId ?? state.getNamespacedToolCallId(toolResult.tool_use_id)
      if (toolResult.is_error) {
        chunks.push({
          type: 'tool-error',
-          toolCallId: toolResult.tool_use_id,
+          toolCallId,
          toolName: pendingCall?.toolName ?? 'unknown',
          input: pendingCall?.input,
          error: toolResult.content,
@@ -343,7 +373,7 @@ function handleUserMessage(
      } else {
        chunks.push({
          type: 'tool-result',
-          toolCallId: toolResult.tool_use_id,
+          toolCallId,
          toolName: pendingCall?.toolName ?? 'unknown',
          input: pendingCall?.input,
          output: toolResult.content,
@@ -457,6 +487,9 @@ function handleStreamEvent(
    }

    case 'message_stop': {
+      if (!state.hasActiveStep()) {
+        break
+      }
      const pending = state.getPendingUsage()
      chunks.push({
        type: 'finish-step',
@@ -514,7 +547,7 @@ function handleContentBlockStart(
    }
    case 'tool_use': {
      const block = state.openToolBlock(index, {
-        toolCallId: contentBlock.id,
+        rawToolCallId: contentBlock.id,
        toolName: contentBlock.name,
        providerMetadata
      })
@@ -111,6 +111,7 @@ const api = {
  setFullScreen: (value: boolean): Promise<void> => ipcRenderer.invoke(IpcChannel.App_SetFullScreen, value),
  isFullScreen: (): Promise<boolean> => ipcRenderer.invoke(IpcChannel.App_IsFullScreen),
  getSystemFonts: (): Promise<string[]> => ipcRenderer.invoke(IpcChannel.App_GetSystemFonts),
+  mockCrashRenderProcess: () => ipcRenderer.invoke(IpcChannel.APP_CrashRenderProcess),
  mac: {
    isProcessTrusted: (): Promise<boolean> => ipcRenderer.invoke(IpcChannel.App_MacIsProcessTrusted),
    requestProcessTrust: (): Promise<boolean> => ipcRenderer.invoke(IpcChannel.App_MacRequestProcessTrust)
@@ -121,7 +122,8 @@ const api = {
  system: {
    getDeviceType: () => ipcRenderer.invoke(IpcChannel.System_GetDeviceType),
    getHostname: () => ipcRenderer.invoke(IpcChannel.System_GetHostname),
-    getCpuName: () => ipcRenderer.invoke(IpcChannel.System_GetCpuName)
+    getCpuName: () => ipcRenderer.invoke(IpcChannel.System_GetCpuName),
+    checkGitBash: (): Promise<boolean> => ipcRenderer.invoke(IpcChannel.System_CheckGitBash)
  },
  devTools: {
    toggle: () => ipcRenderer.invoke(IpcChannel.System_ToggleDevTools)
@@ -32,6 +32,7 @@ import {
  prepareSpecialProviderConfig,
  providerToAiSdkConfig
 } from './provider/providerConfig'
+import type { AiSdkConfig } from './types'

 const logger = loggerService.withContext('ModernAiProvider')

@@ -44,7 +45,7 @@ export type ModernAiProviderConfig = AiSdkMiddlewareConfig & {

 export default class ModernAiProvider {
  private legacyProvider: LegacyAiProvider
-  private config?: ReturnType<typeof providerToAiSdkConfig>
+  private config?: AiSdkConfig
  private actualProvider: Provider
  private model?: Model
  private localProvider: Awaited<AiSdkProvider> | null = null
@@ -89,6 +90,11 @@ export default class ModernAiProvider {
    // 每次请求时重新生成配置以确保API key轮换生效
    this.config = providerToAiSdkConfig(this.actualProvider, this.model)
    logger.debug('Generated provider config for completions', this.config)
+
+    // 检查 config 是否存在
+    if (!this.config) {
+      throw new Error('Provider config is undefined; cannot proceed with completions')
+    }
    if (SUPPORTED_IMAGE_ENDPOINT_LIST.includes(this.config.options.endpoint)) {
      providerConfig.isImageGenerationEndpoint = true
    }
@@ -149,7 +155,8 @@ export default class ModernAiProvider {
    params: StreamTextParams,
    config: ModernAiProviderConfig
  ): Promise<CompletionsResult> {
-    if (config.isImageGenerationEndpoint) {
+    // ai-gateway不是image/generation 端点，所以就先不走legacy了
+    if (config.isImageGenerationEndpoint && config.provider!.id !== SystemProviderIds['ai-gateway']) {
      // 使用 legacy 实现处理图像生成（支持图片编辑等高级功能）
      if (!config.uiMessages) {
        throw new Error('uiMessages is required for image generation endpoint')
@@ -463,8 +470,13 @@ export default class ModernAiProvider {
    // 如果支持新的 AI SDK，使用现代化实现
    if (isModernSdkSupported(this.actualProvider)) {
      try {
+        // 确保 config 已定义
+        if (!this.config) {
+          throw new Error('Provider config is undefined; cannot proceed with generateImage')
+        }
+
        // 确保本地provider已创建
-        if (!this.localProvider) {
+        if (!this.localProvider && this.config) {
          this.localProvider = await createAiSdkProvider(this.config)
          if (!this.localProvider) {
            throw new Error('Local provider not created')
@@ -1,6 +1,6 @@
 import { loggerService } from '@logger'
-import { isNewApiProvider } from '@renderer/config/providers'
 import type { Provider } from '@renderer/types'
+import { isNewApiProvider } from '@renderer/utils/provider'

 import { AihubmixAPIClient } from './aihubmix/AihubmixAPIClient'
 import { AnthropicAPIClient } from './anthropic/AnthropicAPIClient'
@@ -1,12 +1,12 @@
 import { loggerService } from '@logger'
 import {
+  getModelSupportedVerbosity,
  isFunctionCallingModel,
  isNotSupportTemperatureAndTopP,
  isOpenAIModel,
  isSupportFlexServiceTierModel
 } from '@renderer/config/models'
 import { REFERENCE_PROMPT } from '@renderer/config/prompts'
-import { isSupportServiceTierProvider } from '@renderer/config/providers'
 import { getLMStudioKeepAliveTime } from '@renderer/hooks/useLMStudio'
 import { getAssistantSettings } from '@renderer/services/AssistantService'
 import type {
@@ -18,7 +18,6 @@ import type {
  MCPToolResponse,
  MemoryItem,
  Model,
-  OpenAIVerbosity,
  Provider,
  ToolCallResponse,
  WebSearchProviderResponse,
@@ -32,6 +31,7 @@ import {
  OpenAIServiceTiers,
  SystemProviderIds
 } from '@renderer/types'
+import type { OpenAIVerbosity } from '@renderer/types/aiCoreTypes'
 import type { Message } from '@renderer/types/newMessage'
 import type {
  RequestOptions,
@@ -47,6 +47,7 @@ import type {
 import { isJSON, parseJSON } from '@renderer/utils'
 import { addAbortController, removeAbortController } from '@renderer/utils/abortController'
 import { findFileBlocks, getMainTextContent } from '@renderer/utils/messageUtils/find'
+import { isSupportServiceTierProvider } from '@renderer/utils/provider'
 import { defaultTimeout } from '@shared/config/constant'
 import { defaultAppHeaders } from '@shared/utils'
 import { isEmpty } from 'lodash'
@@ -242,12 +243,18 @@ export abstract class BaseApiClient<
    return serviceTierSetting
  }

-  protected getVerbosity(): OpenAIVerbosity {
+  protected getVerbosity(model?: Model): OpenAIVerbosity {
    try {
      const state = window.store?.getState()
      const verbosity = state?.settings?.openAI?.verbosity

      if (verbosity && ['low', 'medium', 'high'].includes(verbosity)) {
+        // If model is provided, check if the verbosity is supported by the model
+        if (model) {
+          const supportedVerbosity = getModelSupportedVerbosity(model)
+          // Use user's verbosity if supported, otherwise use the first supported option
+          return supportedVerbosity.includes(verbosity) ? verbosity : supportedVerbosity[0]
+        }
        return verbosity
      }
    } catch (error) {
@@ -58,10 +58,27 @@ vi.mock('../aws/AwsBedrockAPIClient', () => ({
  AwsBedrockAPIClient: vi.fn().mockImplementation(() => ({}))
 }))

+vi.mock('@renderer/services/AssistantService.ts', () => ({
+  getDefaultAssistant: () => {
+    return {
+      id: 'default',
+      name: 'default',
+      emoji: '😀',
+      prompt: '',
+      topics: [],
+      messages: [],
+      type: 'assistant',
+      regularPhrases: [],
+      settings: {}
+    }
+  }
+}))
+
 // Mock the models config to prevent circular dependency issues
 vi.mock('@renderer/config/models', () => ({
  findTokenLimit: vi.fn(),
  isReasoningModel: vi.fn(),
+  isOpenAILLMModel: vi.fn(),
  SYSTEM_MODELS: {
    silicon: [],
    defaultModel: []
@@ -1,7 +1,8 @@
 import { GoogleGenAI } from '@google/genai'
 import { loggerService } from '@logger'
-import { createVertexProvider, isVertexAIConfigured, isVertexProvider } from '@renderer/hooks/useVertexAI'
+import { createVertexProvider, isVertexAIConfigured } from '@renderer/hooks/useVertexAI'
 import type { Model, Provider, VertexProvider } from '@renderer/types'
+import { isVertexProvider } from '@renderer/utils/provider'
 import { isEmpty } from 'lodash'

 import { AnthropicVertexClient } from '../anthropic/AnthropicVertexClient'
@@ -10,7 +10,6 @@ import { DEFAULT_MAX_TOKENS } from '@renderer/config/constant'
 import {
  findTokenLimit,
  GEMINI_FLASH_MODEL_REGEX,
-  getOpenAIWebSearchParams,
  getThinkModelType,
  isClaudeReasoningModel,
  isDeepSeekHybridInferenceModel,
@@ -35,16 +34,11 @@ import {
  isSupportedThinkingTokenModel,
  isSupportedThinkingTokenQwenModel,
  isSupportedThinkingTokenZhipuModel,
+  isSupportVerbosityModel,
  isVisionModel,
  MODEL_SUPPORTED_REASONING_EFFORT,
  ZHIPU_RESULT_TOKENS
 } from '@renderer/config/models'
-import {
-  isSupportArrayContentProvider,
-  isSupportDeveloperRoleProvider,
-  isSupportEnableThinkingProvider,
-  isSupportStreamOptionsProvider
-} from '@renderer/config/providers'
 import { mapLanguageToQwenMTModel } from '@renderer/config/translate'
 import { processPostsuffixQwen3Model, processReqMessages } from '@renderer/services/ModelMessageService'
 import { estimateTextTokens } from '@renderer/services/TokenService'
@@ -88,6 +82,12 @@ import {
  openAIToolsToMcpTool
 } from '@renderer/utils/mcp-tools'
 import { findFileBlocks, findImageBlocks } from '@renderer/utils/messageUtils/find'
+import {
+  isSupportArrayContentProvider,
+  isSupportDeveloperRoleProvider,
+  isSupportEnableThinkingProvider,
+  isSupportStreamOptionsProvider
+} from '@renderer/utils/provider'
 import { t } from 'i18next'

 import type { GenericChunk } from '../../middleware/schemas'
@@ -733,9 +733,16 @@ export class OpenAIAPIClient extends OpenAIBaseClient<
          ...modalities,
          // groq 有不同的 service tier 配置，不符合 openai 接口类型
          service_tier: this.getServiceTier(model) as OpenAIServiceTier,
+          ...(isSupportVerbosityModel(model)
+            ? {
+                text: {
+                  verbosity: this.getVerbosity(model)
+                }
+              }
+            : {}),
          ...this.getProviderSpecificParameters(assistant, model),
          ...reasoningEffort,
-          ...getOpenAIWebSearchParams(model, enableWebSearch),
+          // ...getOpenAIWebSearchParams(model, enableWebSearch),
          // OpenRouter usage tracking
          ...(this.provider.id === 'openrouter' ? { usage: { include: true } } : {}),
          ...extra_body,
@@ -48,9 +48,8 @@ export abstract class OpenAIBaseClient<
  }

  // 仅适用于openai
-  override getBaseURL(): string {
-    const host = this.provider.apiHost
-    return formatApiHost(host)
+  override getBaseURL(isSupportedAPIVerion: boolean = true): string {
+    return formatApiHost(this.provider.apiHost, isSupportedAPIVerion)
  }

  override async generateImage({
@@ -144,6 +143,11 @@ export abstract class OpenAIBaseClient<
    }

    let apiKeyForSdkInstance = this.apiKey
+    let baseURLForSdkInstance = this.getBaseURL()
+    let headersForSdkInstance = {
+      ...this.defaultHeaders(),
+      ...this.provider.extra_headers
+    }

    if (this.provider.id === 'copilot') {
      const defaultHeaders = store.getState().copilot.defaultHeaders
@@ -151,6 +155,11 @@ export abstract class OpenAIBaseClient<
      // this.provider.apiKey不允许修改
      // this.provider.apiKey = token
      apiKeyForSdkInstance = token
+      baseURLForSdkInstance = this.getBaseURL(false)
+      headersForSdkInstance = {
+        ...headersForSdkInstance,
+        ...COPILOT_DEFAULT_HEADERS
+      }
    }

    if (this.provider.id === 'azure-openai' || this.provider.type === 'azure-openai') {
@@ -164,12 +173,8 @@ export abstract class OpenAIBaseClient<
      this.sdkInstance = new OpenAI({
        dangerouslyAllowBrowser: true,
        apiKey: apiKeyForSdkInstance,
-        baseURL: this.getBaseURL(),
-        defaultHeaders: {
-          ...this.defaultHeaders(),
-          ...this.provider.extra_headers,
-          ...(this.provider.id === 'copilot' ? COPILOT_DEFAULT_HEADERS : {})
-        }
+        baseURL: baseURLForSdkInstance,
+        defaultHeaders: headersForSdkInstance
      }) as TSdkInstance
    }
    return this.sdkInstance
@@ -12,7 +12,6 @@ import {
  isSupportVerbosityModel,
  isVisionModel
 } from '@renderer/config/models'
-import { isSupportDeveloperRoleProvider } from '@renderer/config/providers'
 import { estimateTextTokens } from '@renderer/services/TokenService'
 import type {
  FileMetadata,
@@ -43,6 +42,7 @@ import {
  openAIToolsToMcpTool
 } from '@renderer/utils/mcp-tools'
 import { findFileBlocks, findImageBlocks } from '@renderer/utils/messageUtils/find'
+import { isSupportDeveloperRoleProvider } from '@renderer/utils/provider'
 import { MB } from '@shared/config/constant'
 import { t } from 'i18next'
 import { isEmpty } from 'lodash'
@@ -297,7 +297,31 @@ export class OpenAIResponseAPIClient extends OpenAIBaseClient<

  private convertResponseToMessageContent(response: OpenAI.Responses.Response): ResponseInput {
    const content: OpenAI.Responses.ResponseInput = []
-    content.push(...response.output)
+    response.output.forEach((item) => {
+      if (item.type !== 'apply_patch_call' && item.type !== 'apply_patch_call_output') {
+        content.push(item)
+      } else if (item.type === 'apply_patch_call') {
+        if (item.operation !== undefined) {
+          const applyPatchToolCall: OpenAI.Responses.ResponseInputItem.ApplyPatchCall = {
+            ...item,
+            operation: item.operation
+          }
+          content.push(applyPatchToolCall)
+        } else {
+          logger.warn('Undefined tool call operation for ApplyPatchToolCall.')
+        }
+      } else if (item.type === 'apply_patch_call_output') {
+        if (item.output !== undefined) {
+          const applyPatchToolCallOutput: OpenAI.Responses.ResponseInputItem.ApplyPatchCallOutput = {
+            ...item,
+            output: item.output === null ? undefined : item.output
+          }
+          content.push(applyPatchToolCallOutput)
+        } else {
+          logger.warn('Undefined tool call operation for ApplyPatchToolCall.')
+        }
+      }
+    })
    return content
  }

@@ -496,7 +520,7 @@ export class OpenAIResponseAPIClient extends OpenAIBaseClient<
          ...(isSupportVerbosityModel(model)
            ? {
                text: {
-                  verbosity: this.getVerbosity()
+                  verbosity: this.getVerbosity(model)
                }
              }
            : {}),
@@ -1,6 +1,7 @@
 import { loggerService } from '@logger'
 import { isZhipuModel } from '@renderer/config/models'
 import { getStoreProviders } from '@renderer/hooks/useStore'
+import { getDefaultModel } from '@renderer/services/AssistantService'
 import type { Chunk } from '@renderer/types/chunk'

 import type { CompletionsParams, CompletionsResult } from '../schemas'
@@ -66,7 +67,7 @@ export const ErrorHandlerMiddleware =
  }

 function handleError(error: any, params: CompletionsParams): any {
-  if (isZhipuModel(params.assistant.model) && error.status && !params.enableGenerateImage) {
+  if (isZhipuModel(params.assistant.model || getDefaultModel()) && error.status && !params.enableGenerateImage) {
    return handleZhipuError(error)
  }

@@ -1,10 +1,10 @@
 import type { WebSearchPluginConfig } from '@cherrystudio/ai-core/built-in/plugins'
 import { loggerService } from '@logger'
 import { isSupportedThinkingTokenQwenModel } from '@renderer/config/models'
-import { isSupportEnableThinkingProvider } from '@renderer/config/providers'
 import type { MCPTool } from '@renderer/types'
-import { type Assistant, type Message, type Model, type Provider } from '@renderer/types'
+import { type Assistant, type Message, type Model, type Provider, SystemProviderIds } from '@renderer/types'
 import type { Chunk } from '@renderer/types/chunk'
+import { isSupportEnableThinkingProvider } from '@renderer/utils/provider'
 import type { LanguageModelMiddleware } from 'ai'
 import { extractReasoningMiddleware, simulateStreamingMiddleware } from 'ai'
 import { isEmpty } from 'lodash'
@@ -12,6 +12,7 @@ import { isEmpty } from 'lodash'
 import { isOpenRouterGeminiGenerateImageModel } from '../utils/image'
 import { noThinkMiddleware } from './noThinkMiddleware'
 import { openrouterGenerateImageMiddleware } from './openrouterGenerateImageMiddleware'
+import { openrouterReasoningMiddleware } from './openrouterReasoningMiddleware'
 import { qwenThinkingMiddleware } from './qwenThinkingMiddleware'
 import { toolChoiceMiddleware } from './toolChoiceMiddleware'

@@ -217,6 +218,14 @@ function addProviderSpecificMiddlewares(builder: AiSdkMiddlewareBuilder, config:
      middleware: noThinkMiddleware()
    })
  }
+
+  if (config.provider.id === SystemProviderIds.openrouter && config.enableReasoning) {
+    builder.add({
+      name: 'openrouter-reasoning-redaction',
+      middleware: openrouterReasoningMiddleware()
+    })
+    logger.debug('Added OpenRouter reasoning redaction middleware')
+  }
 }

 /**
@@ -0,0 +1,50 @@
+import type { LanguageModelV2StreamPart } from '@ai-sdk/provider'
+import type { LanguageModelMiddleware } from 'ai'
+
+/**
+ * https://openrouter.ai/docs/docs/best-practices/reasoning-tokens#example-preserving-reasoning-blocks-with-openrouter-and-claude
+ *
+ * @returns LanguageModelMiddleware - a middleware filter redacted block
+ */
+export function openrouterReasoningMiddleware(): LanguageModelMiddleware {
+  const REDACTED_BLOCK = '[REDACTED]'
+  return {
+    middlewareVersion: 'v2',
+    wrapGenerate: async ({ doGenerate }) => {
+      const { content, ...rest } = await doGenerate()
+      const modifiedContent = content.map((part) => {
+        if (part.type === 'reasoning' && part.text.includes(REDACTED_BLOCK)) {
+          return {
+            ...part,
+            text: part.text.replace(REDACTED_BLOCK, '')
+          }
+        }
+        return part
+      })
+      return { content: modifiedContent, ...rest }
+    },
+    wrapStream: async ({ doStream }) => {
+      const { stream, ...rest } = await doStream()
+      return {
+        stream: stream.pipeThrough(
+          new TransformStream<LanguageModelV2StreamPart, LanguageModelV2StreamPart>({
+            transform(
+              chunk: LanguageModelV2StreamPart,
+              controller: TransformStreamDefaultController<LanguageModelV2StreamPart>
+            ) {
+              if (chunk.type === 'reasoning-delta' && chunk.delta.includes(REDACTED_BLOCK)) {
+                controller.enqueue({
+                  ...chunk,
+                  delta: chunk.delta.replace(REDACTED_BLOCK, '')
+                })
+              } else {
+                controller.enqueue(chunk)
+              }
+            }
+          })
+        ),
+        ...rest
+      }
+    }
+  }
+}
@@ -0,0 +1,234 @@
+import type { Message, Model } from '@renderer/types'
+import type { FileMetadata } from '@renderer/types/file'
+import { FileTypes } from '@renderer/types/file'
+import {
+  AssistantMessageStatus,
+  type FileMessageBlock,
+  type ImageMessageBlock,
+  MessageBlockStatus,
+  MessageBlockType,
+  type ThinkingMessageBlock,
+  UserMessageStatus
+} from '@renderer/types/newMessage'
+import { beforeEach, describe, expect, it, vi } from 'vitest'
+
+const { convertFileBlockToFilePartMock, convertFileBlockToTextPartMock } = vi.hoisted(() => ({
+  convertFileBlockToFilePartMock: vi.fn(),
+  convertFileBlockToTextPartMock: vi.fn()
+}))
+
+vi.mock('../fileProcessor', () => ({
+  convertFileBlockToFilePart: convertFileBlockToFilePartMock,
+  convertFileBlockToTextPart: convertFileBlockToTextPartMock
+}))
+
+const visionModelIds = new Set(['gpt-4o-mini', 'qwen-image-edit'])
+const imageEnhancementModelIds = new Set(['qwen-image-edit'])
+
+vi.mock('@renderer/config/models', () => ({
+  isVisionModel: (model: Model) => visionModelIds.has(model.id),
+  isImageEnhancementModel: (model: Model) => imageEnhancementModelIds.has(model.id)
+}))
+
+type MockableMessage = Message & {
+  __mockContent?: string
+  __mockFileBlocks?: FileMessageBlock[]
+  __mockImageBlocks?: ImageMessageBlock[]
+  __mockThinkingBlocks?: ThinkingMessageBlock[]
+}
+
+vi.mock('@renderer/utils/messageUtils/find', () => ({
+  getMainTextContent: (message: Message) => (message as MockableMessage).__mockContent ?? '',
+  findFileBlocks: (message: Message) => (message as MockableMessage).__mockFileBlocks ?? [],
+  findImageBlocks: (message: Message) => (message as MockableMessage).__mockImageBlocks ?? [],
+  findThinkingBlocks: (message: Message) => (message as MockableMessage).__mockThinkingBlocks ?? []
+}))
+
+import { convertMessagesToSdkMessages, convertMessageToSdkParam } from '../messageConverter'
+
+let messageCounter = 0
+let blockCounter = 0
+
+const createModel = (overrides: Partial<Model> = {}): Model => ({
+  id: 'gpt-4o-mini',
+  name: 'GPT-4o mini',
+  provider: 'openai',
+  group: 'openai',
+  ...overrides
+})
+
+const createMessage = (role: Message['role']): MockableMessage =>
+  ({
+    id: `message-${++messageCounter}`,
+    role,
+    assistantId: 'assistant-1',
+    topicId: 'topic-1',
+    createdAt: new Date(2024, 0, 1, 0, 0, messageCounter).toISOString(),
+    status: role === 'assistant' ? AssistantMessageStatus.SUCCESS : UserMessageStatus.SUCCESS,
+    blocks: []
+  }) as MockableMessage
+
+const createFileBlock = (
+  messageId: string,
+  overrides: Partial<Omit<FileMessageBlock, 'file' | 'messageId' | 'type'>> & { file?: Partial<FileMetadata> } = {}
+): FileMessageBlock => {
+  const { file, ...blockOverrides } = overrides
+  const timestamp = new Date(2024, 0, 1, 0, 0, ++blockCounter).toISOString()
+  return {
+    id: blockOverrides.id ?? `file-block-${blockCounter}`,
+    messageId,
+    type: MessageBlockType.FILE,
+    createdAt: blockOverrides.createdAt ?? timestamp,
+    status: blockOverrides.status ?? MessageBlockStatus.SUCCESS,
+    file: {
+      id: file?.id ?? `file-${blockCounter}`,
+      name: file?.name ?? 'document.txt',
+      origin_name: file?.origin_name ?? 'document.txt',
+      path: file?.path ?? '/tmp/document.txt',
+      size: file?.size ?? 1024,
+      ext: file?.ext ?? '.txt',
+      type: file?.type ?? FileTypes.TEXT,
+      created_at: file?.created_at ?? timestamp,
+      count: file?.count ?? 1,
+      ...file
+    },
+    ...blockOverrides
+  }
+}
+
+const createImageBlock = (
+  messageId: string,
+  overrides: Partial<Omit<ImageMessageBlock, 'type' | 'messageId'>> = {}
+): ImageMessageBlock => ({
+  id: overrides.id ?? `image-block-${++blockCounter}`,
+  messageId,
+  type: MessageBlockType.IMAGE,
+  createdAt: overrides.createdAt ?? new Date(2024, 0, 1, 0, 0, blockCounter).toISOString(),
+  status: overrides.status ?? MessageBlockStatus.SUCCESS,
+  url: overrides.url ?? 'https://example.com/image.png',
+  ...overrides
+})
+
+describe('messageConverter', () => {
+  beforeEach(() => {
+    convertFileBlockToFilePartMock.mockReset()
+    convertFileBlockToTextPartMock.mockReset()
+    convertFileBlockToFilePartMock.mockResolvedValue(null)
+    convertFileBlockToTextPartMock.mockResolvedValue(null)
+    messageCounter = 0
+    blockCounter = 0
+  })
+
+  describe('convertMessageToSdkParam', () => {
+    it('includes text and image parts for user messages on vision models', async () => {
+      const model = createModel()
+      const message = createMessage('user')
+      message.__mockContent = 'Describe this picture'
+      message.__mockImageBlocks = [createImageBlock(message.id, { url: 'https://example.com/cat.png' })]
+
+      const result = await convertMessageToSdkParam(message, true, model)
+
+      expect(result).toEqual({
+        role: 'user',
+        content: [
+          { type: 'text', text: 'Describe this picture' },
+          { type: 'image', image: 'https://example.com/cat.png' }
+        ]
+      })
+    })
+
+    it('returns file instructions as a system message when native uploads succeed', async () => {
+      const model = createModel()
+      const message = createMessage('user')
+      message.__mockContent = 'Summarize the PDF'
+      message.__mockFileBlocks = [createFileBlock(message.id)]
+      convertFileBlockToFilePartMock.mockResolvedValueOnce({
+        type: 'file',
+        filename: 'document.pdf',
+        mediaType: 'application/pdf',
+        data: 'fileid://remote-file'
+      })
+
+      const result = await convertMessageToSdkParam(message, false, model)
+
+      expect(result).toEqual([
+        {
+          role: 'system',
+          content: 'fileid://remote-file'
+        },
+        {
+          role: 'user',
+          content: [{ type: 'text', text: 'Summarize the PDF' }]
+        }
+      ])
+    })
+  })
+
+  describe('convertMessagesToSdkMessages', () => {
+    it('appends assistant images to the final user message for image enhancement models', async () => {
+      const model = createModel({ id: 'qwen-image-edit', name: 'Qwen Image Edit', provider: 'qwen', group: 'qwen' })
+      const initialUser = createMessage('user')
+      initialUser.__mockContent = 'Start editing'
+
+      const assistant = createMessage('assistant')
+      assistant.__mockContent = 'Here is the current preview'
+      assistant.__mockImageBlocks = [createImageBlock(assistant.id, { url: 'https://example.com/preview.png' })]
+
+      const finalUser = createMessage('user')
+      finalUser.__mockContent = 'Increase the brightness'
+
+      const result = await convertMessagesToSdkMessages([initialUser, assistant, finalUser], model)
+
+      expect(result).toEqual([
+        {
+          role: 'assistant',
+          content: [{ type: 'text', text: 'Here is the current preview' }]
+        },
+        {
+          role: 'user',
+          content: [
+            { type: 'text', text: 'Increase the brightness' },
+            { type: 'image', image: 'https://example.com/preview.png' }
+          ]
+        }
+      ])
+    })
+
+    it('preserves preceding system instructions when building enhancement payloads', async () => {
+      const model = createModel({ id: 'qwen-image-edit', name: 'Qwen Image Edit', provider: 'qwen', group: 'qwen' })
+      const fileUser = createMessage('user')
+      fileUser.__mockContent = 'Use this document as inspiration'
+      fileUser.__mockFileBlocks = [createFileBlock(fileUser.id, { file: { ext: '.pdf', type: FileTypes.DOCUMENT } })]
+      convertFileBlockToFilePartMock.mockResolvedValueOnce({
+        type: 'file',
+        filename: 'reference.pdf',
+        mediaType: 'application/pdf',
+        data: 'fileid://reference'
+      })
+
+      const assistant = createMessage('assistant')
+      assistant.__mockContent = 'Generated previews ready'
+      assistant.__mockImageBlocks = [createImageBlock(assistant.id, { url: 'https://example.com/reference.png' })]
+
+      const finalUser = createMessage('user')
+      finalUser.__mockContent = 'Apply the edits'
+
+      const result = await convertMessagesToSdkMessages([fileUser, assistant, finalUser], model)
+
+      expect(result).toEqual([
+        { role: 'system', content: 'fileid://reference' },
+        {
+          role: 'assistant',
+          content: [{ type: 'text', text: 'Generated previews ready' }]
+        },
+        {
+          role: 'user',
+          content: [
+            { type: 'text', text: 'Apply the edits' },
+            { type: 'image', image: 'https://example.com/reference.png' }
+          ]
+        }
+      ])
+    })
+  })
+})
@@ -0,0 +1,218 @@
+import type { Assistant, AssistantSettings, Model, Topic } from '@renderer/types'
+import { TopicType } from '@renderer/types'
+import { defaultTimeout } from '@shared/config/constant'
+import { describe, expect, it, vi } from 'vitest'
+
+import { getTemperature, getTimeout, getTopP } from '../modelParameters'
+
+vi.mock('@renderer/services/AssistantService', () => ({
+  getAssistantSettings: (assistant: Assistant): AssistantSettings => ({
+    contextCount: assistant.settings?.contextCount ?? 4096,
+    temperature: assistant.settings?.temperature ?? 0.7,
+    enableTemperature: assistant.settings?.enableTemperature ?? true,
+    topP: assistant.settings?.topP ?? 1,
+    enableTopP: assistant.settings?.enableTopP ?? false,
+    enableMaxTokens: assistant.settings?.enableMaxTokens ?? false,
+    maxTokens: assistant.settings?.maxTokens,
+    streamOutput: assistant.settings?.streamOutput ?? true,
+    toolUseMode: assistant.settings?.toolUseMode ?? 'prompt',
+    defaultModel: assistant.defaultModel,
+    customParameters: assistant.settings?.customParameters ?? [],
+    reasoning_effort: assistant.settings?.reasoning_effort,
+    reasoning_effort_cache: assistant.settings?.reasoning_effort_cache,
+    qwenThinkMode: assistant.settings?.qwenThinkMode
+  })
+}))
+
+vi.mock('@renderer/hooks/useSettings', () => ({
+  getStoreSetting: vi.fn(),
+  useSettings: vi.fn(() => ({})),
+  useNavbarPosition: vi.fn(() => ({ navbarPosition: 'left', isLeftNavbar: true, isTopNavbar: false }))
+}))
+
+vi.mock('@renderer/hooks/useStore', () => ({
+  getStoreProviders: vi.fn(() => [])
+}))
+
+vi.mock('@renderer/store/settings', () => ({
+  default: (state = { settings: {} }) => state
+}))
+
+vi.mock('@renderer/store/assistants', () => ({
+  default: (state = { assistants: [] }) => state
+}))
+
+const createTopic = (assistantId: string): Topic => ({
+  id: `topic-${assistantId}`,
+  assistantId,
+  name: 'topic',
+  createdAt: new Date().toISOString(),
+  updatedAt: new Date().toISOString(),
+  messages: [],
+  type: TopicType.Chat
+})
+
+const createAssistant = (settings: Assistant['settings'] = {}): Assistant => {
+  const assistantId = 'assistant-1'
+  return {
+    id: assistantId,
+    name: 'Test Assistant',
+    prompt: 'prompt',
+    topics: [createTopic(assistantId)],
+    type: 'assistant',
+    settings
+  }
+}
+
+const createModel = (overrides: Partial<Model> = {}): Model => ({
+  id: 'gpt-4o',
+  provider: 'openai',
+  name: 'GPT-4o',
+  group: 'openai',
+  ...overrides
+})
+
+describe('modelParameters', () => {
+  describe('getTemperature', () => {
+    it('returns undefined when reasoning effort is enabled for Claude models', () => {
+      const assistant = createAssistant({ reasoning_effort: 'medium' })
+      const model = createModel({ id: 'claude-opus-4', name: 'Claude Opus 4', provider: 'anthropic', group: 'claude' })
+
+      expect(getTemperature(assistant, model)).toBeUndefined()
+    })
+
+    it('returns undefined for models without temperature/topP support', () => {
+      const assistant = createAssistant({ enableTemperature: true })
+      const model = createModel({ id: 'qwen-mt-large', name: 'Qwen MT', provider: 'qwen', group: 'qwen' })
+
+      expect(getTemperature(assistant, model)).toBeUndefined()
+    })
+
+    it('returns undefined for Claude 4.5 reasoning models when only TopP is enabled', () => {
+      const assistant = createAssistant({ enableTopP: true, enableTemperature: false })
+      const model = createModel({
+        id: 'claude-sonnet-4.5',
+        name: 'Claude Sonnet 4.5',
+        provider: 'anthropic',
+        group: 'claude'
+      })
+
+      expect(getTemperature(assistant, model)).toBeUndefined()
+    })
+
+    it('returns configured temperature when enabled', () => {
+      const assistant = createAssistant({ enableTemperature: true, temperature: 0.42 })
+      const model = createModel({ id: 'gpt-4o', provider: 'openai', group: 'openai' })
+
+      expect(getTemperature(assistant, model)).toBe(0.42)
+    })
+
+    it('returns undefined when temperature is disabled', () => {
+      const assistant = createAssistant({ enableTemperature: false, temperature: 0.9 })
+      const model = createModel({ id: 'gpt-4o', provider: 'openai', group: 'openai' })
+
+      expect(getTemperature(assistant, model)).toBeUndefined()
+    })
+
+    it('clamps temperature to max 1.0 for Zhipu models', () => {
+      const assistant = createAssistant({ enableTemperature: true, temperature: 2.0 })
+      const model = createModel({ id: 'glm-4-plus', name: 'GLM-4 Plus', provider: 'zhipu', group: 'zhipu' })
+
+      expect(getTemperature(assistant, model)).toBe(1.0)
+    })
+
+    it('clamps temperature to max 1.0 for Anthropic models', () => {
+      const assistant = createAssistant({ enableTemperature: true, temperature: 1.5 })
+      const model = createModel({
+        id: 'claude-sonnet-3.5',
+        name: 'Claude 3.5 Sonnet',
+        provider: 'anthropic',
+        group: 'claude'
+      })
+
+      expect(getTemperature(assistant, model)).toBe(1.0)
+    })
+
+    it('clamps temperature to max 1.0 for Moonshot models', () => {
+      const assistant = createAssistant({ enableTemperature: true, temperature: 2.0 })
+      const model = createModel({
+        id: 'moonshot-v1-8k',
+        name: 'Moonshot v1 8k',
+        provider: 'moonshot',
+        group: 'moonshot'
+      })
+
+      expect(getTemperature(assistant, model)).toBe(1.0)
+    })
+
+    it('does not clamp temperature for OpenAI models', () => {
+      const assistant = createAssistant({ enableTemperature: true, temperature: 2.0 })
+      const model = createModel({ id: 'gpt-4o', provider: 'openai', group: 'openai' })
+
+      expect(getTemperature(assistant, model)).toBe(2.0)
+    })
+
+    it('does not clamp temperature when it is already within limits', () => {
+      const assistant = createAssistant({ enableTemperature: true, temperature: 0.8 })
+      const model = createModel({ id: 'glm-4-plus', name: 'GLM-4 Plus', provider: 'zhipu', group: 'zhipu' })
+
+      expect(getTemperature(assistant, model)).toBe(0.8)
+    })
+  })
+
+  describe('getTopP', () => {
+    it('returns undefined when reasoning effort is enabled for Claude models', () => {
+      const assistant = createAssistant({ reasoning_effort: 'high' })
+      const model = createModel({ id: 'claude-opus-4', provider: 'anthropic', group: 'claude' })
+
+      expect(getTopP(assistant, model)).toBeUndefined()
+    })
+
+    it('returns undefined for models without TopP support', () => {
+      const assistant = createAssistant({ enableTopP: true })
+      const model = createModel({ id: 'qwen-mt-small', name: 'Qwen MT', provider: 'qwen', group: 'qwen' })
+
+      expect(getTopP(assistant, model)).toBeUndefined()
+    })
+
+    it('returns undefined for Claude 4.5 reasoning models when temperature is enabled', () => {
+      const assistant = createAssistant({ enableTemperature: true })
+      const model = createModel({
+        id: 'claude-opus-4.5',
+        name: 'Claude Opus 4.5',
+        provider: 'anthropic',
+        group: 'claude'
+      })
+
+      expect(getTopP(assistant, model)).toBeUndefined()
+    })
+
+    it('returns configured TopP when enabled', () => {
+      const assistant = createAssistant({ enableTopP: true, topP: 0.73 })
+      const model = createModel({ id: 'gpt-4o', provider: 'openai', group: 'openai' })
+
+      expect(getTopP(assistant, model)).toBe(0.73)
+    })
+
+    it('returns undefined when TopP is disabled', () => {
+      const assistant = createAssistant({ enableTopP: false, topP: 0.5 })
+      const model = createModel({ id: 'gpt-4o', provider: 'openai', group: 'openai' })
+
+      expect(getTopP(assistant, model)).toBeUndefined()
+    })
+  })
+
+  describe('getTimeout', () => {
+    it('uses an extended timeout for flex service tier models', () => {
+      const model = createModel({ id: 'o3-pro', provider: 'openai', group: 'openai' })
+
+      expect(getTimeout(model)).toBe(15 * 1000 * 60)
+    })
+
+    it('falls back to the default timeout otherwise', () => {
+      const model = createModel({ id: 'gpt-4o', provider: 'openai', group: 'openai' })
+
+      expect(getTimeout(model)).toBe(defaultTimeout)
+    })
+  })
+})
@@ -0,0 +1,31 @@
+import { isClaude4SeriesModel, isClaude45ReasoningModel } from '@renderer/config/models'
+import { getProviderByModel } from '@renderer/services/AssistantService'
+import type { Assistant, Model } from '@renderer/types'
+import { isToolUseModeFunction } from '@renderer/utils/assistant'
+import { isAwsBedrockProvider, isVertexProvider } from '@renderer/utils/provider'
+
+// https://docs.claude.com/en/docs/build-with-claude/extended-thinking#interleaved-thinking
+const INTERLEAVED_THINKING_HEADER = 'interleaved-thinking-2025-05-14'
+// https://docs.claude.com/en/docs/build-with-claude/context-windows#1m-token-context-window
+const CONTEXT_100M_HEADER = 'context-1m-2025-08-07'
+// https://docs.cloud.google.com/vertex-ai/generative-ai/docs/partner-models/claude/web-search
+const WEBSEARCH_HEADER = 'web-search-2025-03-05'
+
+export function addAnthropicHeaders(assistant: Assistant, model: Model): string[] {
+  const anthropicHeaders: string[] = []
+  const provider = getProviderByModel(model)
+  if (
+    isClaude45ReasoningModel(model) &&
+    isToolUseModeFunction(assistant) &&
+    !(isVertexProvider(provider) && isAwsBedrockProvider(provider))
+  ) {
+    anthropicHeaders.push(INTERLEAVED_THINKING_HEADER)
+  }
+  if (isClaude4SeriesModel(model)) {
+    if (isVertexProvider(provider) && assistant.enableWebSearch) {
+      anthropicHeaders.push(WEBSEARCH_HEADER)
+    }
+    anthropicHeaders.push(CONTEXT_100M_HEADER)
+  }
+  return anthropicHeaders
+}
@@ -85,19 +85,6 @@ export function supportsLargeFileUpload(model: Model): boolean {
  })
 }

-/**
- * 检查模型是否支持TopP
- */
-export function supportsTopP(model: Model): boolean {
-  const provider = getProviderByModel(model)
-
-  if (provider?.type === 'anthropic' || model?.endpoint_type === 'anthropic') {
-    return false
-  }
-
-  return true
-}
-
 /**
 * 获取提供商特定的文件大小限制
 */
@@ -3,17 +3,27 @@
 * 处理温度、TopP、超时等基础参数的获取逻辑
 */

+import { DEFAULT_MAX_TOKENS } from '@renderer/config/constant'
 import {
  isClaude45ReasoningModel,
  isClaudeReasoningModel,
+  isMaxTemperatureOneModel,
  isNotSupportTemperatureAndTopP,
-  isSupportedFlexServiceTier
+  isSupportedFlexServiceTier,
+  isSupportedThinkingTokenClaudeModel
 } from '@renderer/config/models'
-import { getAssistantSettings } from '@renderer/services/AssistantService'
+import { getAssistantSettings, getProviderByModel } from '@renderer/services/AssistantService'
 import type { Assistant, Model } from '@renderer/types'
 import { defaultTimeout } from '@shared/config/constant'

+import { getAnthropicThinkingBudget } from '../utils/reasoning'
+
 /**
+ * Claude 4.5 推理模型:
+ * - 只启用 temperature → 使用 temperature
+ * - 只启用 top_p → 使用 top_p
+ * - 同时启用 → temperature 生效,top_p 被忽略
+ * - 都不启用 → 都不使用
 * 获取温度参数
 */
 export function getTemperature(assistant: Assistant, model: Model): number | undefined {
@@ -27,7 +37,11 @@ export function getTemperature(assistant: Assistant, model: Model): number | und
    return undefined
  }
  const assistantSettings = getAssistantSettings(assistant)
-  return assistantSettings?.enableTemperature ? assistantSettings?.temperature : undefined
+  let temperature = assistantSettings?.temperature
+  if (temperature && isMaxTemperatureOneModel(model)) {
+    temperature = Math.min(1, temperature)
+  }
+  return assistantSettings?.enableTemperature ? temperature : undefined
 }

 /**
@@ -56,3 +70,18 @@ export function getTimeout(model: Model): number {
  }
  return defaultTimeout
 }
+
+export function getMaxTokens(assistant: Assistant, model: Model): number | undefined {
+  // NOTE: ai-sdk会把maxToken和budgetToken加起来
+  let { maxTokens = DEFAULT_MAX_TOKENS } = getAssistantSettings(assistant)
+
+  const provider = getProviderByModel(model)
+  if (isSupportedThinkingTokenClaudeModel(model) && ['anthropic', 'aws-bedrock'].includes(provider.type)) {
+    const { reasoning_effort: reasoningEffort } = getAssistantSettings(assistant)
+    const budget = getAnthropicThinkingBudget(maxTokens, reasoningEffort, model.id)
+    if (budget) {
+      maxTokens -= budget
+    }
+  }
+  return maxTokens
+}
@@ -4,22 +4,24 @@
 */

 import { anthropic } from '@ai-sdk/anthropic'
+import { azure } from '@ai-sdk/azure'
 import { google } from '@ai-sdk/google'
 import { vertexAnthropic } from '@ai-sdk/google-vertex/anthropic/edge'
 import { vertex } from '@ai-sdk/google-vertex/edge'
-import type { WebSearchPluginConfig } from '@cherrystudio/ai-core/built-in/plugins'
+import { combineHeaders } from '@ai-sdk/provider-utils'
+import type { AnthropicSearchConfig, WebSearchPluginConfig } from '@cherrystudio/ai-core/built-in/plugins'
 import { isBaseProvider } from '@cherrystudio/ai-core/core/providers/schemas'
 import { loggerService } from '@logger'
 import {
+  isAnthropicModel,
  isGenerateImageModel,
  isOpenRouterBuiltInWebSearchModel,
  isReasoningModel,
  isSupportedReasoningEffortModel,
-  isSupportedThinkingTokenClaudeModel,
  isSupportedThinkingTokenModel,
  isWebSearchModel
 } from '@renderer/config/models'
-import { getAssistantSettings, getDefaultModel } from '@renderer/services/AssistantService'
+import { getDefaultModel } from '@renderer/services/AssistantService'
 import store from '@renderer/store'
 import type { CherryWebSearchConfig } from '@renderer/store/websearch'
 import { type Assistant, type MCPTool, type Provider } from '@renderer/types'
@@ -32,10 +34,9 @@ import { stepCountIs } from 'ai'
 import { getAiSdkProviderId } from '../provider/factory'
 import { setupToolsConfig } from '../utils/mcp'
 import { buildProviderOptions } from '../utils/options'
-import { getAnthropicThinkingBudget } from '../utils/reasoning'
 import { buildProviderBuiltinWebSearchConfig } from '../utils/websearch'
-import { supportsTopP } from './modelCapabilities'
-import { getTemperature, getTopP } from './modelParameters'
+import { addAnthropicHeaders } from './header'
+import { getMaxTokens, getTemperature, getTopP } from './modelParameters'

 const logger = loggerService.withContext('parameterBuilder')

@@ -58,7 +59,7 @@ export async function buildStreamTextParams(
      timeout?: number
      headers?: Record<string, string>
    }
-  } = {}
+  }
 ): Promise<{
  params: StreamTextParams
  modelId: string
@@ -75,8 +76,6 @@ export async function buildStreamTextParams(
  const model = assistant.model || getDefaultModel()
  const aiSdkProviderId = getAiSdkProviderId(provider)

-  let { maxTokens } = getAssistantSettings(assistant)
-
  // 这三个变量透传出来，交给下面启用插件/中间件
  // 也可以在外部构建好再传入buildStreamTextParams
  // FIXME: qwen3即使关闭思考仍然会导致enableReasoning的结果为true
@@ -113,16 +112,6 @@ export async function buildStreamTextParams(
    enableGenerateImage
  })

-  // NOTE: ai-sdk会把maxToken和budgetToken加起来
-  if (
-    enableReasoning &&
-    maxTokens !== undefined &&
-    isSupportedThinkingTokenClaudeModel(model) &&
-    (provider.type === 'anthropic' || provider.type === 'aws-bedrock')
-  ) {
-    maxTokens -= getAnthropicThinkingBudget(assistant, model)
-  }
-
  let webSearchPluginConfig: WebSearchPluginConfig | undefined = undefined
  if (enableWebSearch) {
    if (isBaseProvider(aiSdkProviderId)) {
@@ -139,6 +128,17 @@ export async function buildStreamTextParams(
        maxUses: webSearchConfig.maxResults,
        blockedDomains: blockedDomains.length > 0 ? blockedDomains : undefined
      }) as ProviderDefinedTool
+    } else if (aiSdkProviderId === 'azure-responses') {
+      tools.web_search_preview = azure.tools.webSearchPreview({
+        searchContextSize: webSearchPluginConfig?.openai!.searchContextSize
+      }) as ProviderDefinedTool
+    } else if (aiSdkProviderId === 'azure-anthropic') {
+      const blockedDomains = mapRegexToPatterns(webSearchConfig.excludeDomains)
+      const anthropicSearchOptions: AnthropicSearchConfig = {
+        maxUses: webSearchConfig.maxResults,
+        blockedDomains: blockedDomains.length > 0 ? blockedDomains : undefined
+      }
+      tools.web_search = anthropic.tools.webSearch_20250305(anthropicSearchOptions) as ProviderDefinedTool
    }
  }

@@ -156,9 +156,10 @@ export async function buildStreamTextParams(
        tools.url_context = google.tools.urlContext({}) as ProviderDefinedTool
        break
      case 'anthropic':
+      case 'azure-anthropic':
      case 'google-vertex-anthropic':
        tools.web_fetch = (
-          aiSdkProviderId === 'anthropic'
+          ['anthropic', 'azure-anthropic'].includes(aiSdkProviderId)
            ? anthropic.tools.webFetch_20250910({
                maxUses: webSearchConfig.maxResults,
                blockedDomains: blockedDomains.length > 0 ? blockedDomains : undefined
@@ -172,22 +173,26 @@ export async function buildStreamTextParams(
    }
  }

+  let headers: Record<string, string | undefined> = options.requestOptions?.headers ?? {}
+
+  if (isAnthropicModel(model)) {
+    const newBetaHeaders = { 'anthropic-beta': addAnthropicHeaders(assistant, model).join(',') }
+    headers = combineHeaders(headers, newBetaHeaders)
+  }
+
  // 构建基础参数
  const params: StreamTextParams = {
    messages: sdkMessages,
-    maxOutputTokens: maxTokens,
+    maxOutputTokens: getMaxTokens(assistant, model),
    temperature: getTemperature(assistant, model),
+    topP: getTopP(assistant, model),
    abortSignal: options.requestOptions?.signal,
-    headers: options.requestOptions?.headers,
+    headers,
    providerOptions,
    stopWhen: stepCountIs(20),
    maxRetries: 0
  }

-  if (supportsTopP(model)) {
-    params.topP = getTopP(assistant, model)
-  }
-
  if (tools) {
    params.tools = tools
  }
@@ -23,6 +23,26 @@ vi.mock('@cherrystudio/ai-core', () => ({
  }
 }))

+vi.mock('@renderer/services/AssistantService', () => ({
+  getProviderByModel: vi.fn(),
+  getAssistantSettings: vi.fn(),
+  getDefaultAssistant: vi.fn().mockReturnValue({
+    id: 'default',
+    name: 'Default Assistant',
+    prompt: '',
+    settings: {}
+  })
+}))
+
+vi.mock('@renderer/store/settings', () => ({
+  default: {},
+  settingsSlice: {
+    name: 'settings',
+    reducer: vi.fn(),
+    actions: {}
+  }
+}))
+
 // Mock the provider configs
 vi.mock('../providerConfigs', () => ({
  initializeNewProviders: vi.fn()
@@ -12,7 +12,14 @@ vi.mock('@renderer/services/LoggerService', () => ({
 }))

 vi.mock('@renderer/services/AssistantService', () => ({
-  getProviderByModel: vi.fn()
+  getProviderByModel: vi.fn(),
+  getAssistantSettings: vi.fn(),
+  getDefaultAssistant: vi.fn().mockReturnValue({
+    id: 'default',
+    name: 'Default Assistant',
+    prompt: '',
+    settings: {}
+  })
 }))

 vi.mock('@renderer/store', () => ({
@@ -34,7 +41,7 @@ vi.mock('@renderer/utils/api', () => ({
  }))
 }))

-vi.mock('@renderer/config/providers', async (importOriginal) => {
+vi.mock('@renderer/utils/provider', async (importOriginal) => {
  const actual = (await importOriginal()) as any
  return {
    ...actual,
@@ -53,10 +60,21 @@ vi.mock('@renderer/hooks/useVertexAI', () => ({
  createVertexProvider: vi.fn()
 }))

-import { isCherryAIProvider, isPerplexityProvider } from '@renderer/config/providers'
+vi.mock('@renderer/services/AssistantService', () => ({
+  getProviderByModel: vi.fn(),
+  getAssistantSettings: vi.fn(),
+  getDefaultAssistant: vi.fn().mockReturnValue({
+    id: 'default',
+    name: 'Default Assistant',
+    prompt: '',
+    settings: {}
+  })
+}))
+
 import { getProviderByModel } from '@renderer/services/AssistantService'
 import type { Model, Provider } from '@renderer/types'
 import { formatApiHost } from '@renderer/utils/api'
+import { isCherryAIProvider, isPerplexityProvider } from '@renderer/utils/provider'

 import { COPILOT_DEFAULT_HEADERS, COPILOT_EDITOR_VERSION, isCopilotResponsesModel } from '../constants'
 import { getActualProvider, providerToAiSdkConfig } from '../providerConfig'
@@ -0,0 +1,22 @@
+import type { Provider } from '@renderer/types'
+
+import { provider2Provider, startsWith } from './helper'
+import type { RuleSet } from './types'
+
+// https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry
+const AZURE_ANTHROPIC_RULES: RuleSet = {
+  rules: [
+    {
+      match: startsWith('claude'),
+      provider: (provider: Provider) => ({
+        ...provider,
+        type: 'anthropic',
+        apiHost: provider.apiHost + 'anthropic/v1',
+        id: 'azure-anthropic'
+      })
+    }
+  ],
+  fallbackRule: (provider: Provider) => provider
+}
+
+export const azureAnthropicProviderCreator = provider2Provider.bind(null, AZURE_ANTHROPIC_RULES)
@@ -2,8 +2,10 @@ import { hasProviderConfigByAlias, type ProviderId, resolveProviderConfigId } fr
 import { createProvider as createProviderCore } from '@cherrystudio/ai-core/provider'
 import { loggerService } from '@logger'
 import type { Provider } from '@renderer/types'
+import { isAzureOpenAIProvider, isAzureResponsesEndpoint } from '@renderer/utils/provider'
 import type { Provider as AiSdkProvider } from 'ai'

+import type { AiSdkConfig } from '../types'
 import { initializeNewProviders } from './providerInitialization'

 const logger = loggerService.withContext('ProviderFactory')
@@ -55,9 +57,12 @@ function tryResolveProviderId(identifier: string): ProviderId | null {
 * 获取AI SDK Provider ID
 * 简化版：减少重复逻辑，利用通用解析函数
 */
-export function getAiSdkProviderId(provider: Provider): ProviderId | 'openai-compatible' {
+export function getAiSdkProviderId(provider: Provider): string {
  // 1. 尝试解析provider.id
  const resolvedFromId = tryResolveProviderId(provider.id)
+  if (isAzureOpenAIProvider(provider) && isAzureResponsesEndpoint(provider)) {
+    return 'azure-responses'
+  }
  if (resolvedFromId) {
    return resolvedFromId
  }
@@ -73,11 +78,11 @@ export function getAiSdkProviderId(provider: Provider): ProviderId | 'openai-com
  if (provider.apiHost.includes('api.openai.com')) {
    return 'openai-chat'
  }
-  // 3. 最后的fallback（通常会成为openai-compatible）
-  return provider.id as ProviderId
+  // 3. 最后的fallback（使用provider本身的id）
+  return provider.id
 }

-export async function createAiSdkProvider(config) {
+export async function createAiSdkProvider(config: AiSdkConfig): Promise<AiSdkProvider | null> {
  let localProvider: Awaited<AiSdkProvider> | null = null
  try {
    if (config.providerId === 'openai' && config.options?.mode === 'chat') {
@@ -1,19 +1,5 @@
-import {
-  formatPrivateKey,
-  hasProviderConfig,
-  ProviderConfigFactory,
-  type ProviderId,
-  type ProviderSettingsMap
-} from '@cherrystudio/ai-core/provider'
+import { formatPrivateKey, hasProviderConfig, ProviderConfigFactory } from '@cherrystudio/ai-core/provider'
 import { isOpenAIChatCompletionOnlyModel } from '@renderer/config/models'
-import {
-  isAnthropicProvider,
-  isAzureOpenAIProvider,
-  isCherryAIProvider,
-  isGeminiProvider,
-  isNewApiProvider,
-  isPerplexityProvider
-} from '@renderer/config/providers'
 import {
  getAwsBedrockAccessKeyId,
  getAwsBedrockApiKey,
@@ -21,14 +7,25 @@ import {
  getAwsBedrockRegion,
  getAwsBedrockSecretAccessKey
 } from '@renderer/hooks/useAwsBedrock'
-import { createVertexProvider, isVertexAIConfigured, isVertexProvider } from '@renderer/hooks/useVertexAI'
+import { createVertexProvider, isVertexAIConfigured } from '@renderer/hooks/useVertexAI'
 import { getProviderByModel } from '@renderer/services/AssistantService'
 import store from '@renderer/store'
 import { isSystemProvider, type Model, type Provider, SystemProviderIds } from '@renderer/types'
 import { formatApiHost, formatAzureOpenAIApiHost, formatVertexApiHost, routeToEndpoint } from '@renderer/utils/api'
+import {
+  isAnthropicProvider,
+  isAzureOpenAIProvider,
+  isCherryAIProvider,
+  isGeminiProvider,
+  isNewApiProvider,
+  isPerplexityProvider,
+  isVertexProvider
+} from '@renderer/utils/provider'
 import { cloneDeep } from 'lodash'

+import type { AiSdkConfig } from '../types'
 import { aihubmixProviderCreator, newApiResolverCreator, vertexAnthropicProviderCreator } from './config'
+import { azureAnthropicProviderCreator } from './config/azure-anthropic'
 import { COPILOT_DEFAULT_HEADERS } from './constants'
 import { getAiSdkProviderId } from './factory'

@@ -74,6 +71,9 @@ function handleSpecialProviders(model: Model, provider: Provider): Provider {
      return vertexAnthropicProviderCreator(model, provider)
    }
  }
+  if (isAzureOpenAIProvider(provider)) {
+    return azureAnthropicProviderCreator(model, provider)
+  }
  return provider
 }

@@ -131,13 +131,7 @@ export function getActualProvider(model: Model): Provider {
 * 将 Provider 配置转换为新 AI SDK 格式
 * 简化版：利用新的别名映射系统
 */
-export function providerToAiSdkConfig(
-  actualProvider: Provider,
-  model: Model
-): {
-  providerId: ProviderId | 'openai-compatible'
-  options: ProviderSettingsMap[keyof ProviderSettingsMap]
-} {
+export function providerToAiSdkConfig(actualProvider: Provider, model: Model): AiSdkConfig {
  const aiSdkProviderId = getAiSdkProviderId(actualProvider)

  // 构建基础配置
@@ -191,13 +185,10 @@ export function providerToAiSdkConfig(
  // azure
  // https://learn.microsoft.com/en-us/azure/ai-foundry/openai/latest
  // https://learn.microsoft.com/en-us/azure/ai-foundry/openai/how-to/responses?tabs=python-key#responses-api
-  if (aiSdkProviderId === 'azure' || actualProvider.type === 'azure-openai') {
-    // extraOptions.apiVersion = actualProvider.apiVersion === 'preview' ? 'v1' : actualProvider.apiVersion 默认使用v1，不使用azure endpoint
-    if (actualProvider.apiVersion === 'preview' || actualProvider.apiVersion === 'v1') {
-      extraOptions.mode = 'responses'
-    } else {
-      extraOptions.mode = 'chat'
-    }
+  if (aiSdkProviderId === 'azure-responses') {
+    extraOptions.mode = 'responses'
+  } else if (aiSdkProviderId === 'azure') {
+    extraOptions.mode = 'chat'
  }

  // bedrock
@@ -227,10 +218,17 @@ export function providerToAiSdkConfig(
    baseConfig.baseURL += aiSdkProviderId === 'google-vertex' ? '/publishers/google' : '/publishers/anthropic/models'
  }

+  // cherryin
+  if (aiSdkProviderId === 'cherryin') {
+    if (model.endpoint_type) {
+      extraOptions.endpointType = model.endpoint_type
+    }
+  }
+
  if (hasProviderConfig(aiSdkProviderId) && aiSdkProviderId !== 'openai-compatible') {
    const options = ProviderConfigFactory.fromProvider(aiSdkProviderId, baseConfig, extraOptions)
    return {
-      providerId: aiSdkProviderId as ProviderId,
+      providerId: aiSdkProviderId,
      options
    }
  }
@@ -32,6 +32,14 @@ export const NEW_PROVIDER_CONFIGS: ProviderConfig[] = [
    supportsImageGeneration: true,
    aliases: ['vertexai-anthropic']
  },
+  {
+    id: 'azure-anthropic',
+    name: 'Azure AI Anthropic',
+    import: () => import('@ai-sdk/anthropic'),
+    creatorFunctionName: 'createAnthropic',
+    supportsImageGeneration: false,
+    aliases: ['azure-anthropic']
+  },
  {
    id: 'github-copilot-openai-compatible',
    name: 'GitHub Copilot OpenAI Compatible',
@@ -133,7 +133,7 @@ export class AiSdkSpanAdapter {

    // 详细记录转换过程
    const operationId = attributes['ai.operationId']
-    logger.info('Converting AI SDK span to SpanEntity', {
+    logger.debug('Converting AI SDK span to SpanEntity', {
      spanName: spanName,
      operationId,
      spanTag,
@@ -149,7 +149,7 @@ export class AiSdkSpanAdapter {
    })

    if (tokenUsage) {
-      logger.info('Token usage data found', {
+      logger.debug('Token usage data found', {
        spanName: spanName,
        operationId,
        usage: tokenUsage,
@@ -158,7 +158,7 @@ export class AiSdkSpanAdapter {
    }

    if (inputs || outputs) {
-      logger.info('Input/Output data extracted', {
+      logger.debug('Input/Output data extracted', {
        spanName: spanName,
        operationId,
        hasInputs: !!inputs,
@@ -170,7 +170,7 @@ export class AiSdkSpanAdapter {
    }

    if (Object.keys(typeSpecificData).length > 0) {
-      logger.info('Type-specific data extracted', {
+      logger.debug('Type-specific data extracted', {
        spanName: spanName,
        operationId,
        typeSpecificKeys: Object.keys(typeSpecificData),
@@ -204,7 +204,7 @@ export class AiSdkSpanAdapter {
      modelName: modelName || this.extractModelFromAttributes(attributes)
    }

-    logger.info('AI SDK span successfully converted to SpanEntity', {
+    logger.debug('AI SDK span successfully converted to SpanEntity', {
      spanName: spanName,
      operationId,
      spanId: spanContext.spanId,
@@ -0,0 +1,15 @@
+/**
+ * This type definition file is only for renderer.
+ * It cannot be migrated to @renderer/types since files within it are actually being used by both main and renderer.
+ * If we do that, main would throw an error because it cannot import a module which imports a type from a browser-enviroment-only package.
+ * (ai-core package is set as browser-enviroment-only)
+ *
+ * TODO: We should separate them clearly. Keep renderer only types in renderer, and main only types in main, and shared types in shared.
+ */
+
+import type { ProviderSettingsMap } from '@cherrystudio/ai-core/provider'
+
+export type AiSdkConfig = {
+  providerId: string
+  options: ProviderSettingsMap[keyof ProviderSettingsMap]
+}
@@ -0,0 +1,121 @@
+/**
+ * image.ts Unit Tests
+ * Tests for Gemini image generation utilities
+ */
+
+import type { Model, Provider } from '@renderer/types'
+import { SystemProviderIds } from '@renderer/types'
+import { describe, expect, it } from 'vitest'
+
+import { buildGeminiGenerateImageParams, isOpenRouterGeminiGenerateImageModel } from '../image'
+
+describe('image utils', () => {
+  describe('buildGeminiGenerateImageParams', () => {
+    it('should return correct response modalities', () => {
+      const result = buildGeminiGenerateImageParams()
+
+      expect(result).toEqual({
+        responseModalities: ['TEXT', 'IMAGE']
+      })
+    })
+
+    it('should return an object with responseModalities property', () => {
+      const result = buildGeminiGenerateImageParams()
+
+      expect(result).toHaveProperty('responseModalities')
+      expect(Array.isArray(result.responseModalities)).toBe(true)
+      expect(result.responseModalities).toHaveLength(2)
+    })
+  })
+
+  describe('isOpenRouterGeminiGenerateImageModel', () => {
+    const mockOpenRouterProvider: Provider = {
+      id: SystemProviderIds.openrouter,
+      name: 'OpenRouter',
+      apiKey: 'test-key',
+      apiHost: 'https://openrouter.ai/api/v1',
+      isSystem: true
+    } as Provider
+
+    const mockOtherProvider: Provider = {
+      id: SystemProviderIds.openai,
+      name: 'OpenAI',
+      apiKey: 'test-key',
+      apiHost: 'https://api.openai.com/v1',
+      isSystem: true
+    } as Provider
+
+    it('should return true for OpenRouter Gemini 2.5 Flash Image model', () => {
+      const model: Model = {
+        id: 'google/gemini-2.5-flash-image-preview',
+        name: 'Gemini 2.5 Flash Image',
+        provider: SystemProviderIds.openrouter
+      } as Model
+
+      const result = isOpenRouterGeminiGenerateImageModel(model, mockOpenRouterProvider)
+      expect(result).toBe(true)
+    })
+
+    it('should return false for non-Gemini model on OpenRouter', () => {
+      const model: Model = {
+        id: 'openai/gpt-4',
+        name: 'GPT-4',
+        provider: SystemProviderIds.openrouter
+      } as Model
+
+      const result = isOpenRouterGeminiGenerateImageModel(model, mockOpenRouterProvider)
+      expect(result).toBe(false)
+    })
+
+    it('should return false for Gemini model on non-OpenRouter provider', () => {
+      const model: Model = {
+        id: 'gemini-2.5-flash-image-preview',
+        name: 'Gemini 2.5 Flash Image',
+        provider: SystemProviderIds.gemini
+      } as Model
+
+      const result = isOpenRouterGeminiGenerateImageModel(model, mockOtherProvider)
+      expect(result).toBe(false)
+    })
+
+    it('should return false for Gemini model without image suffix', () => {
+      const model: Model = {
+        id: 'google/gemini-2.5-flash',
+        name: 'Gemini 2.5 Flash',
+        provider: SystemProviderIds.openrouter
+      } as Model
+
+      const result = isOpenRouterGeminiGenerateImageModel(model, mockOpenRouterProvider)
+      expect(result).toBe(false)
+    })
+
+    it('should handle model ID with partial match', () => {
+      const model: Model = {
+        id: 'google/gemini-2.5-flash-image-generation',
+        name: 'Gemini Image Gen',
+        provider: SystemProviderIds.openrouter
+      } as Model
+
+      const result = isOpenRouterGeminiGenerateImageModel(model, mockOpenRouterProvider)
+      expect(result).toBe(true)
+    })
+
+    it('should return false for custom provider', () => {
+      const customProvider: Provider = {
+        id: 'custom-provider-123',
+        name: 'Custom Provider',
+        apiKey: 'test-key',
+        apiHost: 'https://custom.com'
+      } as Provider
+
+      const model: Model = {
+        id: 'gemini-2.5-flash-image-preview',
+        name: 'Gemini 2.5 Flash Image',
+        provider: 'custom-provider-123'
+      } as Model
+
+      const result = isOpenRouterGeminiGenerateImageModel(model, customProvider)
+      expect(result).toBe(false)
+    })
+  })
+})
@@ -0,0 +1,435 @@
+/**
+ * mcp.ts Unit Tests
+ * Tests for MCP tools configuration and conversion utilities
+ */
+
+import type { MCPTool } from '@renderer/types'
+import type { Tool } from 'ai'
+import { beforeEach, describe, expect, it, vi } from 'vitest'
+
+import { convertMcpToolsToAiSdkTools, setupToolsConfig } from '../mcp'
+
+// Mock dependencies
+vi.mock('@logger', () => ({
+  loggerService: {
+    withContext: () => ({
+      debug: vi.fn(),
+      error: vi.fn(),
+      warn: vi.fn(),
+      info: vi.fn()
+    })
+  }
+}))
+
+vi.mock('@renderer/utils/mcp-tools', () => ({
+  getMcpServerByTool: vi.fn(() => ({ id: 'test-server', autoApprove: false })),
+  isToolAutoApproved: vi.fn(() => false),
+  callMCPTool: vi.fn(async () => ({
+    content: [{ type: 'text', text: 'Tool executed successfully' }],
+    isError: false
+  }))
+}))
+
+vi.mock('@renderer/utils/userConfirmation', () => ({
+  requestToolConfirmation: vi.fn(async () => true)
+}))
+
+describe('mcp utils', () => {
+  beforeEach(() => {
+    vi.clearAllMocks()
+  })
+
+  describe('setupToolsConfig', () => {
+    it('should return undefined when no MCP tools provided', () => {
+      const result = setupToolsConfig()
+      expect(result).toBeUndefined()
+    })
+
+    it('should return undefined when empty MCP tools array provided', () => {
+      const result = setupToolsConfig([])
+      expect(result).toBeUndefined()
+    })
+
+    it('should convert MCP tools to AI SDK tools format', () => {
+      const mcpTools: MCPTool[] = [
+        {
+          id: 'test-tool-1',
+          serverId: 'test-server',
+          serverName: 'test-server',
+          name: 'test-tool',
+          description: 'A test tool',
+          type: 'mcp',
+          inputSchema: {
+            type: 'object',
+            properties: {
+              query: { type: 'string' }
+            }
+          }
+        }
+      ]
+
+      const result = setupToolsConfig(mcpTools)
+
+      expect(result).not.toBeUndefined()
+      expect(Object.keys(result!)).toEqual(['test-tool'])
+      expect(result!['test-tool']).toHaveProperty('description')
+      expect(result!['test-tool']).toHaveProperty('inputSchema')
+      expect(result!['test-tool']).toHaveProperty('execute')
+    })
+
+    it('should handle multiple MCP tools', () => {
+      const mcpTools: MCPTool[] = [
+        {
+          id: 'tool1-id',
+          serverId: 'server1',
+          serverName: 'server1',
+          name: 'tool1',
+          description: 'First tool',
+          type: 'mcp',
+          inputSchema: {
+            type: 'object',
+            properties: {}
+          }
+        },
+        {
+          id: 'tool2-id',
+          serverId: 'server2',
+          serverName: 'server2',
+          name: 'tool2',
+          description: 'Second tool',
+          type: 'mcp',
+          inputSchema: {
+            type: 'object',
+            properties: {}
+          }
+        }
+      ]
+
+      const result = setupToolsConfig(mcpTools)
+
+      expect(result).not.toBeUndefined()
+      expect(Object.keys(result!)).toHaveLength(2)
+      expect(Object.keys(result!)).toEqual(['tool1', 'tool2'])
+    })
+  })
+
+  describe('convertMcpToolsToAiSdkTools', () => {
+    it('should convert single MCP tool to AI SDK tool', () => {
+      const mcpTools: MCPTool[] = [
+        {
+          id: 'get-weather-id',
+          serverId: 'weather-server',
+          serverName: 'weather-server',
+          name: 'get-weather',
+          description: 'Get weather information',
+          type: 'mcp',
+          inputSchema: {
+            type: 'object',
+            properties: {
+              location: { type: 'string' }
+            },
+            required: ['location']
+          }
+        }
+      ]
+
+      const result = convertMcpToolsToAiSdkTools(mcpTools)
+
+      expect(Object.keys(result)).toEqual(['get-weather'])
+
+      const tool = result['get-weather'] as Tool
+      expect(tool.description).toBe('Get weather information')
+      expect(tool.inputSchema).toBeDefined()
+      expect(typeof tool.execute).toBe('function')
+    })
+
+    it('should handle tool without description', () => {
+      const mcpTools: MCPTool[] = [
+        {
+          id: 'no-desc-tool-id',
+          serverId: 'test-server',
+          serverName: 'test-server',
+          name: 'no-desc-tool',
+          type: 'mcp',
+          inputSchema: {
+            type: 'object',
+            properties: {}
+          }
+        }
+      ]
+
+      const result = convertMcpToolsToAiSdkTools(mcpTools)
+
+      expect(Object.keys(result)).toEqual(['no-desc-tool'])
+      const tool = result['no-desc-tool'] as Tool
+      expect(tool.description).toBe('Tool from test-server')
+    })
+
+    it('should convert empty tools array', () => {
+      const result = convertMcpToolsToAiSdkTools([])
+      expect(result).toEqual({})
+    })
+
+    it('should handle complex input schemas', () => {
+      const mcpTools: MCPTool[] = [
+        {
+          id: 'complex-tool-id',
+          serverId: 'server',
+          serverName: 'server',
+          name: 'complex-tool',
+          description: 'Tool with complex schema',
+          type: 'mcp',
+          inputSchema: {
+            type: 'object',
+            properties: {
+              name: { type: 'string' },
+              age: { type: 'number' },
+              tags: {
+                type: 'array',
+                items: { type: 'string' }
+              },
+              metadata: {
+                type: 'object',
+                properties: {
+                  key: { type: 'string' }
+                }
+              }
+            },
+            required: ['name']
+          }
+        }
+      ]
+
+      const result = convertMcpToolsToAiSdkTools(mcpTools)
+
+      expect(Object.keys(result)).toEqual(['complex-tool'])
+      const tool = result['complex-tool'] as Tool
+      expect(tool.inputSchema).toBeDefined()
+      expect(typeof tool.execute).toBe('function')
+    })
+
+    it('should preserve tool names with special characters', () => {
+      const mcpTools: MCPTool[] = [
+        {
+          id: 'special-tool-id',
+          serverId: 'server',
+          serverName: 'server',
+          name: 'tool_with-special.chars',
+          description: 'Special chars tool',
+          type: 'mcp',
+          inputSchema: {
+            type: 'object',
+            properties: {}
+          }
+        }
+      ]
+
+      const result = convertMcpToolsToAiSdkTools(mcpTools)
+      expect(Object.keys(result)).toEqual(['tool_with-special.chars'])
+    })
+
+    it('should handle multiple tools with different schemas', () => {
+      const mcpTools: MCPTool[] = [
+        {
+          id: 'string-tool-id',
+          serverId: 'server1',
+          serverName: 'server1',
+          name: 'string-tool',
+          description: 'String tool',
+          type: 'mcp',
+          inputSchema: {
+            type: 'object',
+            properties: {
+              input: { type: 'string' }
+            }
+          }
+        },
+        {
+          id: 'number-tool-id',
+          serverId: 'server2',
+          serverName: 'server2',
+          name: 'number-tool',
+          description: 'Number tool',
+          type: 'mcp',
+          inputSchema: {
+            type: 'object',
+            properties: {
+              count: { type: 'number' }
+            }
+          }
+        },
+        {
+          id: 'boolean-tool-id',
+          serverId: 'server3',
+          serverName: 'server3',
+          name: 'boolean-tool',
+          description: 'Boolean tool',
+          type: 'mcp',
+          inputSchema: {
+            type: 'object',
+            properties: {
+              enabled: { type: 'boolean' }
+            }
+          }
+        }
+      ]
+
+      const result = convertMcpToolsToAiSdkTools(mcpTools)
+
+      expect(Object.keys(result).sort()).toEqual(['boolean-tool', 'number-tool', 'string-tool'])
+      expect(result['string-tool']).toBeDefined()
+      expect(result['number-tool']).toBeDefined()
+      expect(result['boolean-tool']).toBeDefined()
+    })
+  })
+
+  describe('tool execution', () => {
+    it('should execute tool with user confirmation', async () => {
+      const { callMCPTool } = await import('@renderer/utils/mcp-tools')
+      const { requestToolConfirmation } = await import('@renderer/utils/userConfirmation')
+
+      vi.mocked(requestToolConfirmation).mockResolvedValue(true)
+      vi.mocked(callMCPTool).mockResolvedValue({
+        content: [{ type: 'text', text: 'Success' }],
+        isError: false
+      })
+
+      const mcpTools: MCPTool[] = [
+        {
+          id: 'test-exec-tool-id',
+          serverId: 'test-server',
+          serverName: 'test-server',
+          name: 'test-exec-tool',
+          description: 'Test execution tool',
+          type: 'mcp',
+          inputSchema: {
+            type: 'object',
+            properties: {}
+          }
+        }
+      ]
+
+      const tools = convertMcpToolsToAiSdkTools(mcpTools)
+      const tool = tools['test-exec-tool'] as Tool
+      const result = await tool.execute!({}, { messages: [], abortSignal: undefined, toolCallId: 'test-call-123' })
+
+      expect(requestToolConfirmation).toHaveBeenCalled()
+      expect(callMCPTool).toHaveBeenCalled()
+      expect(result).toEqual({
+        content: [{ type: 'text', text: 'Success' }],
+        isError: false
+      })
+    })
+
+    it('should handle user cancellation', async () => {
+      const { requestToolConfirmation } = await import('@renderer/utils/userConfirmation')
+      const { callMCPTool } = await import('@renderer/utils/mcp-tools')
+
+      vi.mocked(requestToolConfirmation).mockResolvedValue(false)
+
+      const mcpTools: MCPTool[] = [
+        {
+          id: 'cancelled-tool-id',
+          serverId: 'test-server',
+          serverName: 'test-server',
+          name: 'cancelled-tool',
+          description: 'Tool to cancel',
+          type: 'mcp',
+          inputSchema: {
+            type: 'object',
+            properties: {}
+          }
+        }
+      ]
+
+      const tools = convertMcpToolsToAiSdkTools(mcpTools)
+      const tool = tools['cancelled-tool'] as Tool
+      const result = await tool.execute!({}, { messages: [], abortSignal: undefined, toolCallId: 'cancel-call-123' })
+
+      expect(requestToolConfirmation).toHaveBeenCalled()
+      expect(callMCPTool).not.toHaveBeenCalled()
+      expect(result).toEqual({
+        content: [
+          {
+            type: 'text',
+            text: 'User declined to execute tool "cancelled-tool".'
+          }
+        ],
+        isError: false
+      })
+    })
+
+    it('should handle tool execution error', async () => {
+      const { callMCPTool } = await import('@renderer/utils/mcp-tools')
+      const { requestToolConfirmation } = await import('@renderer/utils/userConfirmation')
+
+      vi.mocked(requestToolConfirmation).mockResolvedValue(true)
+      vi.mocked(callMCPTool).mockResolvedValue({
+        content: [{ type: 'text', text: 'Error occurred' }],
+        isError: true
+      })
+
+      const mcpTools: MCPTool[] = [
+        {
+          id: 'error-tool-id',
+          serverId: 'test-server',
+          serverName: 'test-server',
+          name: 'error-tool',
+          description: 'Tool that errors',
+          type: 'mcp',
+          inputSchema: {
+            type: 'object',
+            properties: {}
+          }
+        }
+      ]
+
+      const tools = convertMcpToolsToAiSdkTools(mcpTools)
+      const tool = tools['error-tool'] as Tool
+
+      await expect(
+        tool.execute!({}, { messages: [], abortSignal: undefined, toolCallId: 'error-call-123' })
+      ).rejects.toEqual({
+        content: [{ type: 'text', text: 'Error occurred' }],
+        isError: true
+      })
+    })
+
+    it('should auto-approve when enabled', async () => {
+      const { callMCPTool, isToolAutoApproved } = await import('@renderer/utils/mcp-tools')
+      const { requestToolConfirmation } = await import('@renderer/utils/userConfirmation')
+
+      vi.mocked(isToolAutoApproved).mockReturnValue(true)
+      vi.mocked(callMCPTool).mockResolvedValue({
+        content: [{ type: 'text', text: 'Auto-approved success' }],
+        isError: false
+      })
+
+      const mcpTools: MCPTool[] = [
+        {
+          id: 'auto-approve-tool-id',
+          serverId: 'test-server',
+          serverName: 'test-server',
+          name: 'auto-approve-tool',
+          description: 'Auto-approved tool',
+          type: 'mcp',
+          inputSchema: {
+            type: 'object',
+            properties: {}
+          }
+        }
+      ]
+
+      const tools = convertMcpToolsToAiSdkTools(mcpTools)
+      const tool = tools['auto-approve-tool'] as Tool
+      const result = await tool.execute!({}, { messages: [], abortSignal: undefined, toolCallId: 'auto-call-123' })
+
+      expect(requestToolConfirmation).not.toHaveBeenCalled()
+      expect(callMCPTool).toHaveBeenCalled()
+      expect(result).toEqual({
+        content: [{ type: 'text', text: 'Auto-approved success' }],
+        isError: false
+      })
+    })
+  })
+})
@@ -0,0 +1,545 @@
+/**
+ * options.ts Unit Tests
+ * Tests for building provider-specific options
+ */
+
+import type { Assistant, Model, Provider } from '@renderer/types'
+import { OpenAIServiceTiers, SystemProviderIds } from '@renderer/types'
+import { beforeEach, describe, expect, it, vi } from 'vitest'
+
+import { buildProviderOptions } from '../options'
+
+// Mock dependencies
+vi.mock('@cherrystudio/ai-core/provider', async (importOriginal) => {
+  const actual = (await importOriginal()) as object
+  return {
+    ...actual,
+    baseProviderIdSchema: {
+      safeParse: vi.fn((id) => {
+        const baseProviders = [
+          'openai',
+          'openai-chat',
+          'azure',
+          'azure-responses',
+          'huggingface',
+          'anthropic',
+          'google',
+          'xai',
+          'deepseek',
+          'openrouter',
+          'openai-compatible'
+        ]
+        if (baseProviders.includes(id)) {
+          return { success: true, data: id }
+        }
+        return { success: false }
+      })
+    },
+    customProviderIdSchema: {
+      safeParse: vi.fn((id) => {
+        const customProviders = ['google-vertex', 'google-vertex-anthropic', 'bedrock']
+        if (customProviders.includes(id)) {
+          return { success: true, data: id }
+        }
+        return { success: false, error: new Error('Invalid provider') }
+      })
+    }
+  }
+})
+
+vi.mock('../provider/factory', () => ({
+  getAiSdkProviderId: vi.fn((provider) => {
+    // Simulate the provider ID mapping
+    const mapping: Record<string, string> = {
+      [SystemProviderIds.gemini]: 'google',
+      [SystemProviderIds.openai]: 'openai',
+      [SystemProviderIds.anthropic]: 'anthropic',
+      [SystemProviderIds.grok]: 'xai',
+      [SystemProviderIds.deepseek]: 'deepseek',
+      [SystemProviderIds.openrouter]: 'openrouter'
+    }
+    return mapping[provider.id] || provider.id
+  })
+}))
+
+vi.mock('@renderer/config/models', async (importOriginal) => ({
+  ...(await importOriginal()),
+  isOpenAIModel: vi.fn((model) => model.id.includes('gpt') || model.id.includes('o1')),
+  isQwenMTModel: vi.fn(() => false),
+  isSupportFlexServiceTierModel: vi.fn(() => true),
+  isOpenAILLMModel: vi.fn(() => true),
+  SYSTEM_MODELS: {
+    defaultModel: [
+      { id: 'default-1', name: 'Default 1' },
+      { id: 'default-2', name: 'Default 2' },
+      { id: 'default-3', name: 'Default 3' }
+    ]
+  }
+}))
+
+vi.mock(import('@renderer/utils/provider'), async (importOriginal) => {
+  return {
+    ...(await importOriginal()),
+    isSupportServiceTierProvider: vi.fn((provider) => {
+      return [SystemProviderIds.openai, SystemProviderIds.groq].includes(provider.id)
+    })
+  }
+})
+
+vi.mock('@renderer/store/settings', () => ({
+  default: (state = { settings: {} }) => state
+}))
+
+vi.mock('@renderer/hooks/useSettings', () => ({
+  getStoreSetting: vi.fn((key) => {
+    if (key === 'openAI') {
+      return { summaryText: 'off', verbosity: 'medium' } as any
+    }
+    return {}
+  })
+}))
+
+vi.mock('@renderer/services/AssistantService', () => ({
+  getDefaultAssistant: vi.fn(() => ({
+    id: 'default',
+    name: 'Default Assistant',
+    settings: {}
+  })),
+  getAssistantSettings: vi.fn(() => ({
+    reasoning_effort: 'medium',
+    maxTokens: 4096
+  })),
+  getProviderByModel: vi.fn((model: Model) => ({
+    id: model.provider,
+    name: 'Mock Provider'
+  }))
+}))
+
+vi.mock('../reasoning', () => ({
+  getOpenAIReasoningParams: vi.fn(() => ({ reasoningEffort: 'medium' })),
+  getAnthropicReasoningParams: vi.fn(() => ({
+    thinking: { type: 'enabled', budgetTokens: 5000 }
+  })),
+  getGeminiReasoningParams: vi.fn(() => ({
+    thinkingConfig: { include_thoughts: true }
+  })),
+  getXAIReasoningParams: vi.fn(() => ({ reasoningEffort: 'high' })),
+  getBedrockReasoningParams: vi.fn(() => ({
+    reasoningConfig: { type: 'enabled', budgetTokens: 5000 }
+  })),
+  getReasoningEffort: vi.fn(() => ({ reasoningEffort: 'medium' })),
+  getCustomParameters: vi.fn(() => ({}))
+}))
+
+vi.mock('../image', () => ({
+  buildGeminiGenerateImageParams: vi.fn(() => ({
+    responseModalities: ['TEXT', 'IMAGE']
+  }))
+}))
+
+vi.mock('../websearch', () => ({
+  getWebSearchParams: vi.fn(() => ({ enable_search: true }))
+}))
+
+const ensureWindowApi = () => {
+  const globalWindow = window as any
+  globalWindow.api = globalWindow.api || {}
+  globalWindow.api.getAppInfo = globalWindow.api.getAppInfo || vi.fn(async () => ({ notesPath: '' }))
+}
+
+ensureWindowApi()
+
+describe('options utils', () => {
+  const mockAssistant: Assistant = {
+    id: 'test-assistant',
+    name: 'Test Assistant',
+    settings: {}
+  } as Assistant
+
+  const mockModel: Model = {
+    id: 'gpt-4',
+    name: 'GPT-4',
+    provider: SystemProviderIds.openai
+  } as Model
+
+  beforeEach(() => {
+    vi.clearAllMocks()
+  })
+
+  describe('buildProviderOptions', () => {
+    describe('OpenAI provider', () => {
+      const openaiProvider: Provider = {
+        id: SystemProviderIds.openai,
+        name: 'OpenAI',
+        type: 'openai-response',
+        apiKey: 'test-key',
+        apiHost: 'https://api.openai.com/v1',
+        isSystem: true
+      } as Provider
+
+      it('should build basic OpenAI options', () => {
+        const result = buildProviderOptions(mockAssistant, mockModel, openaiProvider, {
+          enableReasoning: false,
+          enableWebSearch: false,
+          enableGenerateImage: false
+        })
+
+        expect(result).toHaveProperty('openai')
+        expect(result.openai).toBeDefined()
+      })
+
+      it('should include reasoning parameters when enabled', () => {
+        const result = buildProviderOptions(mockAssistant, mockModel, openaiProvider, {
+          enableReasoning: true,
+          enableWebSearch: false,
+          enableGenerateImage: false
+        })
+
+        expect(result.openai).toHaveProperty('reasoningEffort')
+        expect(result.openai.reasoningEffort).toBe('medium')
+      })
+
+      it('should include service tier when supported', () => {
+        const providerWithServiceTier: Provider = {
+          ...openaiProvider,
+          serviceTier: OpenAIServiceTiers.auto
+        }
+
+        const result = buildProviderOptions(mockAssistant, mockModel, providerWithServiceTier, {
+          enableReasoning: false,
+          enableWebSearch: false,
+          enableGenerateImage: false
+        })
+
+        expect(result.openai).toHaveProperty('serviceTier')
+        expect(result.openai.serviceTier).toBe(OpenAIServiceTiers.auto)
+      })
+    })
+
+    describe('Anthropic provider', () => {
+      const anthropicProvider: Provider = {
+        id: SystemProviderIds.anthropic,
+        name: 'Anthropic',
+        type: 'anthropic',
+        apiKey: 'test-key',
+        apiHost: 'https://api.anthropic.com',
+        isSystem: true
+      } as Provider
+
+      const anthropicModel: Model = {
+        id: 'claude-3-5-sonnet-20241022',
+        name: 'Claude 3.5 Sonnet',
+        provider: SystemProviderIds.anthropic
+      } as Model
+
+      it('should build basic Anthropic options', () => {
+        const result = buildProviderOptions(mockAssistant, anthropicModel, anthropicProvider, {
+          enableReasoning: false,
+          enableWebSearch: false,
+          enableGenerateImage: false
+        })
+
+        expect(result).toHaveProperty('anthropic')
+        expect(result.anthropic).toBeDefined()
+      })
+
+      it('should include reasoning parameters when enabled', () => {
+        const result = buildProviderOptions(mockAssistant, anthropicModel, anthropicProvider, {
+          enableReasoning: true,
+          enableWebSearch: false,
+          enableGenerateImage: false
+        })
+
+        expect(result.anthropic).toHaveProperty('thinking')
+        expect(result.anthropic.thinking).toEqual({
+          type: 'enabled',
+          budgetTokens: 5000
+        })
+      })
+    })
+
+    describe('Google provider', () => {
+      const googleProvider: Provider = {
+        id: SystemProviderIds.gemini,
+        name: 'Google',
+        type: 'gemini',
+        apiKey: 'test-key',
+        apiHost: 'https://generativelanguage.googleapis.com',
+        isSystem: true,
+        models: [{ id: 'gemini-2.0-flash-exp' }] as Model[]
+      } as Provider
+
+      const googleModel: Model = {
+        id: 'gemini-2.0-flash-exp',
+        name: 'Gemini 2.0 Flash',
+        provider: SystemProviderIds.gemini
+      } as Model
+
+      it('should build basic Google options', () => {
+        const result = buildProviderOptions(mockAssistant, googleModel, googleProvider, {
+          enableReasoning: false,
+          enableWebSearch: false,
+          enableGenerateImage: false
+        })
+
+        expect(result).toHaveProperty('google')
+        expect(result.google).toBeDefined()
+      })
+
+      it('should include reasoning parameters when enabled', () => {
+        const result = buildProviderOptions(mockAssistant, googleModel, googleProvider, {
+          enableReasoning: true,
+          enableWebSearch: false,
+          enableGenerateImage: false
+        })
+
+        expect(result.google).toHaveProperty('thinkingConfig')
+        expect(result.google.thinkingConfig).toEqual({
+          include_thoughts: true
+        })
+      })
+
+      it('should include image generation parameters when enabled', () => {
+        const result = buildProviderOptions(mockAssistant, googleModel, googleProvider, {
+          enableReasoning: false,
+          enableWebSearch: false,
+          enableGenerateImage: true
+        })
+
+        expect(result.google).toHaveProperty('responseModalities')
+        expect(result.google.responseModalities).toEqual(['TEXT', 'IMAGE'])
+      })
+    })
+
+    describe('xAI provider', () => {
+      const xaiProvider = {
+        id: SystemProviderIds.grok,
+        name: 'xAI',
+        type: 'new-api',
+        apiKey: 'test-key',
+        apiHost: 'https://api.x.ai/v1',
+        isSystem: true,
+        models: [] as Model[]
+      } as Provider
+
+      const xaiModel: Model = {
+        id: 'grok-2-latest',
+        name: 'Grok 2',
+        provider: SystemProviderIds.grok
+      } as Model
+
+      it('should build basic xAI options', () => {
+        const result = buildProviderOptions(mockAssistant, xaiModel, xaiProvider, {
+          enableReasoning: false,
+          enableWebSearch: false,
+          enableGenerateImage: false
+        })
+
+        expect(result).toHaveProperty('xai')
+        expect(result.xai).toBeDefined()
+      })
+
+      it('should include reasoning parameters when enabled', () => {
+        const result = buildProviderOptions(mockAssistant, xaiModel, xaiProvider, {
+          enableReasoning: true,
+          enableWebSearch: false,
+          enableGenerateImage: false
+        })
+
+        expect(result.xai).toHaveProperty('reasoningEffort')
+        expect(result.xai.reasoningEffort).toBe('high')
+      })
+    })
+
+    describe('DeepSeek provider', () => {
+      const deepseekProvider: Provider = {
+        id: SystemProviderIds.deepseek,
+        name: 'DeepSeek',
+        type: 'openai',
+        apiKey: 'test-key',
+        apiHost: 'https://api.deepseek.com',
+        isSystem: true
+      } as Provider
+
+      const deepseekModel: Model = {
+        id: 'deepseek-chat',
+        name: 'DeepSeek Chat',
+        provider: SystemProviderIds.deepseek
+      } as Model
+
+      it('should build basic DeepSeek options', () => {
+        const result = buildProviderOptions(mockAssistant, deepseekModel, deepseekProvider, {
+          enableReasoning: false,
+          enableWebSearch: false,
+          enableGenerateImage: false
+        })
+
+        expect(result).toHaveProperty('deepseek')
+        expect(result.deepseek).toBeDefined()
+      })
+    })
+
+    describe('OpenRouter provider', () => {
+      const openrouterProvider: Provider = {
+        id: SystemProviderIds.openrouter,
+        name: 'OpenRouter',
+        type: 'openai',
+        apiKey: 'test-key',
+        apiHost: 'https://openrouter.ai/api/v1',
+        isSystem: true
+      } as Provider
+
+      const openrouterModel: Model = {
+        id: 'openai/gpt-4',
+        name: 'GPT-4',
+        provider: SystemProviderIds.openrouter
+      } as Model
+
+      it('should build basic OpenRouter options', () => {
+        const result = buildProviderOptions(mockAssistant, openrouterModel, openrouterProvider, {
+          enableReasoning: false,
+          enableWebSearch: false,
+          enableGenerateImage: false
+        })
+
+        expect(result).toHaveProperty('openrouter')
+        expect(result.openrouter).toBeDefined()
+      })
+
+      it('should include web search parameters when enabled', () => {
+        const result = buildProviderOptions(mockAssistant, openrouterModel, openrouterProvider, {
+          enableReasoning: false,
+          enableWebSearch: true,
+          enableGenerateImage: false
+        })
+
+        expect(result.openrouter).toHaveProperty('enable_search')
+      })
+    })
+
+    describe('Custom parameters', () => {
+      it('should merge custom parameters', async () => {
+        const { getCustomParameters } = await import('../reasoning')
+
+        vi.mocked(getCustomParameters).mockReturnValue({
+          custom_param: 'custom_value',
+          another_param: 123
+        })
+
+        const result = buildProviderOptions(
+          mockAssistant,
+          mockModel,
+          {
+            id: SystemProviderIds.openai,
+            name: 'OpenAI',
+            type: 'openai',
+            apiKey: 'test-key',
+            apiHost: 'https://api.openai.com/v1'
+          } as Provider,
+          {
+            enableReasoning: false,
+            enableWebSearch: false,
+            enableGenerateImage: false
+          }
+        )
+
+        expect(result.openai).toHaveProperty('custom_param')
+        expect(result.openai.custom_param).toBe('custom_value')
+        expect(result.openai).toHaveProperty('another_param')
+        expect(result.openai.another_param).toBe(123)
+      })
+    })
+
+    describe('Multiple capabilities', () => {
+      const googleProvider = {
+        id: SystemProviderIds.gemini,
+        name: 'Google',
+        type: 'gemini',
+        apiKey: 'test-key',
+        apiHost: 'https://generativelanguage.googleapis.com',
+        isSystem: true,
+        models: [] as Model[]
+      } as Provider
+
+      const googleModel: Model = {
+        id: 'gemini-2.0-flash-exp',
+        name: 'Gemini 2.0 Flash',
+        provider: SystemProviderIds.gemini
+      } as Model
+
+      it('should combine reasoning and image generation', () => {
+        const result = buildProviderOptions(mockAssistant, googleModel, googleProvider, {
+          enableReasoning: true,
+          enableWebSearch: false,
+          enableGenerateImage: true
+        })
+
+        expect(result.google).toHaveProperty('thinkingConfig')
+        expect(result.google).toHaveProperty('responseModalities')
+      })
+
+      it('should handle all capabilities enabled', () => {
+        const result = buildProviderOptions(mockAssistant, googleModel, googleProvider, {
+          enableReasoning: true,
+          enableWebSearch: true,
+          enableGenerateImage: true
+        })
+
+        expect(result.google).toBeDefined()
+        expect(Object.keys(result.google).length).toBeGreaterThan(0)
+      })
+    })
+
+    describe('Vertex AI providers', () => {
+      it('should map google-vertex to google', () => {
+        const vertexProvider = {
+          id: 'google-vertex',
+          name: 'Vertex AI',
+          type: 'vertexai',
+          apiKey: 'test-key',
+          apiHost: 'https://vertex-ai.googleapis.com',
+          models: [] as Model[]
+        } as Provider
+
+        const vertexModel: Model = {
+          id: 'gemini-2.0-flash-exp',
+          name: 'Gemini 2.0 Flash',
+          provider: 'google-vertex'
+        } as Model
+
+        const result = buildProviderOptions(mockAssistant, vertexModel, vertexProvider, {
+          enableReasoning: false,
+          enableWebSearch: false,
+          enableGenerateImage: false
+        })
+
+        expect(result).toHaveProperty('google')
+      })
+
+      it('should map google-vertex-anthropic to anthropic', () => {
+        const vertexAnthropicProvider = {
+          id: 'google-vertex-anthropic',
+          name: 'Vertex AI Anthropic',
+          type: 'vertex-anthropic',
+          apiKey: 'test-key',
+          apiHost: 'https://vertex-ai.googleapis.com',
+          models: [] as Model[]
+        } as Provider
+
+        const vertexModel: Model = {
+          id: 'claude-3-5-sonnet-20241022',
+          name: 'Claude 3.5 Sonnet',
+          provider: 'google-vertex-anthropic'
+        } as Model
+
+        const result = buildProviderOptions(mockAssistant, vertexModel, vertexAnthropicProvider, {
+          enableReasoning: false,
+          enableWebSearch: false,
+          enableGenerateImage: false
+        })
+
+        expect(result).toHaveProperty('anthropic')
+      })
+    })
+  })
+})
@@ -0,0 +1,967 @@
+/**
+ * reasoning.ts Unit Tests
+ * Tests for reasoning parameter generation utilities
+ */
+
+import { getStoreSetting } from '@renderer/hooks/useSettings'
+import type { SettingsState } from '@renderer/store/settings'
+import type { Assistant, Model, Provider } from '@renderer/types'
+import { SystemProviderIds } from '@renderer/types'
+import { beforeEach, describe, expect, it, vi } from 'vitest'
+
+import {
+  getAnthropicReasoningParams,
+  getBedrockReasoningParams,
+  getCustomParameters,
+  getGeminiReasoningParams,
+  getOpenAIReasoningParams,
+  getReasoningEffort,
+  getXAIReasoningParams
+} from '../reasoning'
+
+function defaultGetStoreSetting<K extends keyof SettingsState>(key: K): SettingsState[K] {
+  if (key === 'openAI') {
+    return {
+      summaryText: 'auto',
+      verbosity: 'medium'
+    } as SettingsState[K]
+  }
+  return undefined as SettingsState[K]
+}
+
+// Mock dependencies
+vi.mock('@logger', () => ({
+  loggerService: {
+    withContext: () => ({
+      debug: vi.fn(),
+      error: vi.fn(),
+      warn: vi.fn(),
+      info: vi.fn()
+    })
+  }
+}))
+
+vi.mock('@renderer/store/settings', () => ({
+  default: (state = { settings: {} }) => state
+}))
+
+vi.mock('@renderer/store/llm', () => ({
+  initialState: {},
+  default: (state = { llm: {} }) => state
+}))
+
+vi.mock('@renderer/config/constant', () => ({
+  DEFAULT_MAX_TOKENS: 4096,
+  isMac: false,
+  isWin: false,
+  TOKENFLUX_HOST: 'mock-host'
+}))
+
+vi.mock('@renderer/utils/provider', () => ({
+  isSupportEnableThinkingProvider: vi.fn((provider) => {
+    return [SystemProviderIds.dashscope, SystemProviderIds.silicon].includes(provider.id)
+  })
+}))
+
+vi.mock('@renderer/config/models', async (importOriginal) => {
+  const actual: any = await importOriginal()
+  return {
+    ...actual,
+    isReasoningModel: vi.fn(() => false),
+    isOpenAIDeepResearchModel: vi.fn(() => false),
+    isOpenAIModel: vi.fn(() => false),
+    isSupportedReasoningEffortOpenAIModel: vi.fn(() => false),
+    isSupportedThinkingTokenQwenModel: vi.fn(() => false),
+    isQwenReasoningModel: vi.fn(() => false),
+    isSupportedThinkingTokenClaudeModel: vi.fn(() => false),
+    isSupportedThinkingTokenGeminiModel: vi.fn(() => false),
+    isSupportedThinkingTokenDoubaoModel: vi.fn(() => false),
+    isSupportedThinkingTokenZhipuModel: vi.fn(() => false),
+    isSupportedReasoningEffortModel: vi.fn(() => false),
+    isDeepSeekHybridInferenceModel: vi.fn(() => false),
+    isSupportedReasoningEffortGrokModel: vi.fn(() => false),
+    getThinkModelType: vi.fn(() => 'default'),
+    isDoubaoSeedAfter251015: vi.fn(() => false),
+    isDoubaoThinkingAutoModel: vi.fn(() => false),
+    isGrok4FastReasoningModel: vi.fn(() => false),
+    isGrokReasoningModel: vi.fn(() => false),
+    isOpenAIReasoningModel: vi.fn(() => false),
+    isQwenAlwaysThinkModel: vi.fn(() => false),
+    isSupportedThinkingTokenHunyuanModel: vi.fn(() => false),
+    isSupportedThinkingTokenModel: vi.fn(() => false),
+    isGPT51SeriesModel: vi.fn(() => false)
+  }
+})
+
+vi.mock('@renderer/hooks/useSettings', () => ({
+  getStoreSetting: vi.fn(defaultGetStoreSetting)
+}))
+
+vi.mock('@renderer/services/AssistantService', () => ({
+  getAssistantSettings: vi.fn((assistant) => ({
+    maxTokens: assistant?.settings?.maxTokens || 4096,
+    reasoning_effort: assistant?.settings?.reasoning_effort
+  })),
+  getProviderByModel: vi.fn((model) => ({
+    id: model.provider,
+    name: 'Test Provider'
+  })),
+  getDefaultAssistant: vi.fn(() => ({
+    id: 'default',
+    name: 'Default Assistant',
+    settings: {}
+  }))
+}))
+
+const ensureWindowApi = () => {
+  const globalWindow = window as any
+  globalWindow.api = globalWindow.api || {}
+  globalWindow.api.getAppInfo = globalWindow.api.getAppInfo || vi.fn(async () => ({ notesPath: '' }))
+}
+
+ensureWindowApi()
+
+describe('reasoning utils', () => {
+  beforeEach(() => {
+    vi.resetAllMocks()
+  })
+
+  describe('getReasoningEffort', () => {
+    it('should return empty object for non-reasoning model', async () => {
+      const model: Model = {
+        id: 'gpt-4',
+        name: 'GPT-4',
+        provider: SystemProviderIds.openai
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {}
+      } as Assistant
+
+      const result = getReasoningEffort(assistant, model)
+      expect(result).toEqual({})
+    })
+
+    it('should disable reasoning for OpenRouter when no reasoning effort set', async () => {
+      const { isReasoningModel } = await import('@renderer/config/models')
+
+      vi.mocked(isReasoningModel).mockReturnValue(true)
+
+      const model: Model = {
+        id: 'anthropic/claude-sonnet-4',
+        name: 'Claude Sonnet 4',
+        provider: SystemProviderIds.openrouter
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {}
+      } as Assistant
+
+      const result = getReasoningEffort(assistant, model)
+      expect(result).toEqual({ reasoning: { enabled: false, exclude: true } })
+    })
+
+    it('should handle Qwen models with enable_thinking', async () => {
+      const { isReasoningModel, isSupportedThinkingTokenQwenModel, isQwenReasoningModel } = await import(
+        '@renderer/config/models'
+      )
+
+      vi.mocked(isReasoningModel).mockReturnValue(true)
+      vi.mocked(isSupportedThinkingTokenQwenModel).mockReturnValue(true)
+      vi.mocked(isQwenReasoningModel).mockReturnValue(true)
+
+      const model: Model = {
+        id: 'qwen-plus',
+        name: 'Qwen Plus',
+        provider: SystemProviderIds.dashscope
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {
+          reasoning_effort: 'medium'
+        }
+      } as Assistant
+
+      const result = getReasoningEffort(assistant, model)
+      expect(result).toHaveProperty('enable_thinking')
+    })
+
+    it('should handle Claude models with thinking config', async () => {
+      const {
+        isSupportedThinkingTokenClaudeModel,
+        isReasoningModel,
+        isQwenReasoningModel,
+        isSupportedThinkingTokenGeminiModel,
+        isSupportedThinkingTokenDoubaoModel,
+        isSupportedThinkingTokenZhipuModel,
+        isSupportedReasoningEffortModel
+      } = await import('@renderer/config/models')
+
+      vi.mocked(isReasoningModel).mockReturnValue(true)
+      vi.mocked(isSupportedThinkingTokenClaudeModel).mockReturnValue(true)
+      vi.mocked(isQwenReasoningModel).mockReturnValue(false)
+      vi.mocked(isSupportedThinkingTokenGeminiModel).mockReturnValue(false)
+      vi.mocked(isSupportedThinkingTokenDoubaoModel).mockReturnValue(false)
+      vi.mocked(isSupportedThinkingTokenZhipuModel).mockReturnValue(false)
+      vi.mocked(isSupportedReasoningEffortModel).mockReturnValue(false)
+
+      const model: Model = {
+        id: 'claude-3-7-sonnet',
+        name: 'Claude 3.7 Sonnet',
+        provider: SystemProviderIds.anthropic
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {
+          reasoning_effort: 'high',
+          maxTokens: 4096
+        }
+      } as Assistant
+
+      const result = getReasoningEffort(assistant, model)
+      expect(result).toEqual({
+        thinking: {
+          type: 'enabled',
+          budget_tokens: expect.any(Number)
+        }
+      })
+    })
+
+    it('should handle Gemini Flash models with thinking budget 0', async () => {
+      const {
+        isSupportedThinkingTokenGeminiModel,
+        isReasoningModel,
+        isQwenReasoningModel,
+        isSupportedThinkingTokenClaudeModel,
+        isSupportedThinkingTokenDoubaoModel,
+        isSupportedThinkingTokenZhipuModel,
+        isOpenAIDeepResearchModel,
+        isSupportedThinkingTokenQwenModel,
+        isSupportedThinkingTokenHunyuanModel,
+        isDeepSeekHybridInferenceModel
+      } = await import('@renderer/config/models')
+
+      vi.mocked(isReasoningModel).mockReturnValue(true)
+      vi.mocked(isOpenAIDeepResearchModel).mockReturnValue(false)
+      vi.mocked(isSupportedThinkingTokenGeminiModel).mockReturnValue(true)
+      vi.mocked(isQwenReasoningModel).mockReturnValue(false)
+      vi.mocked(isSupportedThinkingTokenClaudeModel).mockReturnValue(false)
+      vi.mocked(isSupportedThinkingTokenDoubaoModel).mockReturnValue(false)
+      vi.mocked(isSupportedThinkingTokenZhipuModel).mockReturnValue(false)
+      vi.mocked(isSupportedThinkingTokenQwenModel).mockReturnValue(false)
+      vi.mocked(isSupportedThinkingTokenHunyuanModel).mockReturnValue(false)
+      vi.mocked(isDeepSeekHybridInferenceModel).mockReturnValue(false)
+
+      const model: Model = {
+        id: 'gemini-2.5-flash',
+        name: 'Gemini 2.5 Flash',
+        provider: SystemProviderIds.openai
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {}
+      } as Assistant
+
+      const result = getReasoningEffort(assistant, model)
+      expect(result).toEqual({
+        extra_body: {
+          google: {
+            thinking_config: {
+              thinking_budget: 0
+            }
+          }
+        }
+      })
+    })
+
+    it('should handle GPT-5.1 reasoning model with effort levels', async () => {
+      const {
+        isReasoningModel,
+        isOpenAIDeepResearchModel,
+        isSupportedReasoningEffortModel,
+        isGPT51SeriesModel,
+        getThinkModelType
+      } = await import('@renderer/config/models')
+
+      vi.mocked(isReasoningModel).mockReturnValue(true)
+      vi.mocked(isOpenAIDeepResearchModel).mockReturnValue(false)
+      vi.mocked(isSupportedReasoningEffortModel).mockReturnValue(true)
+      vi.mocked(getThinkModelType).mockReturnValue('gpt5_1')
+      vi.mocked(isGPT51SeriesModel).mockReturnValue(true)
+
+      const model: Model = {
+        id: 'gpt-5.1',
+        name: 'GPT-5.1',
+        provider: SystemProviderIds.openai
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {
+          reasoning_effort: 'none'
+        }
+      } as Assistant
+
+      const result = getReasoningEffort(assistant, model)
+      expect(result).toEqual({
+        reasoningEffort: 'none'
+      })
+    })
+
+    it('should handle DeepSeek hybrid inference models', async () => {
+      const { isReasoningModel, isDeepSeekHybridInferenceModel } = await import('@renderer/config/models')
+
+      vi.mocked(isReasoningModel).mockReturnValue(true)
+      vi.mocked(isDeepSeekHybridInferenceModel).mockReturnValue(true)
+
+      const model: Model = {
+        id: 'deepseek-v3.1',
+        name: 'DeepSeek V3.1',
+        provider: SystemProviderIds.silicon
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {
+          reasoning_effort: 'high'
+        }
+      } as Assistant
+
+      const result = getReasoningEffort(assistant, model)
+      expect(result).toEqual({
+        enable_thinking: true
+      })
+    })
+
+    it('should return medium effort for deep research models', async () => {
+      const { isReasoningModel, isOpenAIDeepResearchModel } = await import('@renderer/config/models')
+
+      vi.mocked(isReasoningModel).mockReturnValue(true)
+      vi.mocked(isOpenAIDeepResearchModel).mockReturnValue(true)
+
+      const model: Model = {
+        id: 'o3-deep-research',
+        provider: SystemProviderIds.openai
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {}
+      } as Assistant
+
+      const result = getReasoningEffort(assistant, model)
+      expect(result).toEqual({ reasoning_effort: 'medium' })
+    })
+
+    it('should return empty for groq provider', async () => {
+      const { getProviderByModel } = await import('@renderer/services/AssistantService')
+
+      vi.mocked(getProviderByModel).mockReturnValue({
+        id: 'groq',
+        name: 'Groq'
+      } as Provider)
+
+      const model: Model = {
+        id: 'groq-model',
+        name: 'Groq Model',
+        provider: 'groq'
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {}
+      } as Assistant
+
+      const result = getReasoningEffort(assistant, model)
+      expect(result).toEqual({})
+    })
+  })
+
+  describe('getOpenAIReasoningParams', () => {
+    it('should return empty object for non-reasoning model', async () => {
+      const model: Model = {
+        id: 'gpt-4',
+        name: 'GPT-4',
+        provider: SystemProviderIds.openai
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {}
+      } as Assistant
+
+      const result = getOpenAIReasoningParams(assistant, model)
+      expect(result).toEqual({})
+    })
+
+    it('should return empty when no reasoning effort set', async () => {
+      const model: Model = {
+        id: 'o1-preview',
+        name: 'O1 Preview',
+        provider: SystemProviderIds.openai
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {}
+      } as Assistant
+
+      const result = getOpenAIReasoningParams(assistant, model)
+      expect(result).toEqual({})
+    })
+
+    it('should return reasoning effort for OpenAI models', async () => {
+      const { isReasoningModel, isOpenAIModel, isSupportedReasoningEffortOpenAIModel } = await import(
+        '@renderer/config/models'
+      )
+
+      vi.mocked(isReasoningModel).mockReturnValue(true)
+      vi.mocked(isOpenAIModel).mockReturnValue(true)
+      vi.mocked(isSupportedReasoningEffortOpenAIModel).mockReturnValue(true)
+
+      const model: Model = {
+        id: 'gpt-5.1',
+        name: 'GPT 5.1',
+        provider: SystemProviderIds.openai
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {
+          reasoning_effort: 'high'
+        }
+      } as Assistant
+
+      const result = getOpenAIReasoningParams(assistant, model)
+      expect(result).toEqual({
+        reasoningEffort: 'high',
+        reasoningSummary: 'auto'
+      })
+    })
+
+    it('should include reasoning summary when not o1-pro', async () => {
+      const { isReasoningModel, isOpenAIModel, isSupportedReasoningEffortOpenAIModel } = await import(
+        '@renderer/config/models'
+      )
+
+      vi.mocked(isReasoningModel).mockReturnValue(true)
+      vi.mocked(isOpenAIModel).mockReturnValue(true)
+      vi.mocked(isSupportedReasoningEffortOpenAIModel).mockReturnValue(true)
+
+      const model: Model = {
+        id: 'gpt-5',
+        provider: SystemProviderIds.openai
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {
+          reasoning_effort: 'medium'
+        }
+      } as Assistant
+
+      const result = getOpenAIReasoningParams(assistant, model)
+      expect(result).toEqual({
+        reasoningEffort: 'medium',
+        reasoningSummary: 'auto'
+      })
+    })
+
+    it('should not include reasoning summary for o1-pro', async () => {
+      const { isReasoningModel, isOpenAIDeepResearchModel, isSupportedReasoningEffortOpenAIModel } = await import(
+        '@renderer/config/models'
+      )
+
+      vi.mocked(isReasoningModel).mockReturnValue(true)
+      vi.mocked(isOpenAIDeepResearchModel).mockReturnValue(false)
+      vi.mocked(isSupportedReasoningEffortOpenAIModel).mockReturnValue(true)
+      vi.mocked(getStoreSetting).mockReturnValue({ summaryText: 'off' } as any)
+
+      const model: Model = {
+        id: 'o1-pro',
+        name: 'O1 Pro',
+        provider: SystemProviderIds.openai
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {
+          reasoning_effort: 'high'
+        }
+      } as Assistant
+
+      const result = getOpenAIReasoningParams(assistant, model)
+      expect(result).toEqual({
+        reasoningEffort: 'high',
+        reasoningSummary: undefined
+      })
+    })
+
+    it('should force medium effort for deep research models', async () => {
+      const { isReasoningModel, isOpenAIModel, isOpenAIDeepResearchModel, isSupportedReasoningEffortOpenAIModel } =
+        await import('@renderer/config/models')
+      const { getStoreSetting } = await import('@renderer/hooks/useSettings')
+
+      vi.mocked(isReasoningModel).mockReturnValue(true)
+      vi.mocked(isOpenAIModel).mockReturnValue(true)
+      vi.mocked(isOpenAIDeepResearchModel).mockReturnValue(true)
+      vi.mocked(isSupportedReasoningEffortOpenAIModel).mockReturnValue(true)
+      vi.mocked(getStoreSetting).mockReturnValue({ summaryText: 'off' } as any)
+
+      const model: Model = {
+        id: 'o3-deep-research',
+        name: 'O3 Mini',
+        provider: SystemProviderIds.openai
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {
+          reasoning_effort: 'high'
+        }
+      } as Assistant
+
+      const result = getOpenAIReasoningParams(assistant, model)
+      expect(result).toEqual({
+        reasoningEffort: 'medium',
+        reasoningSummary: 'off'
+      })
+    })
+  })
+
+  describe('getAnthropicReasoningParams', () => {
+    it('should return empty for non-reasoning model', async () => {
+      const { isReasoningModel } = await import('@renderer/config/models')
+
+      vi.mocked(isReasoningModel).mockReturnValue(false)
+
+      const model: Model = {
+        id: 'claude-3-5-sonnet',
+        name: 'Claude 3.5 Sonnet',
+        provider: SystemProviderIds.anthropic
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {}
+      } as Assistant
+
+      const result = getAnthropicReasoningParams(assistant, model)
+      expect(result).toEqual({})
+    })
+
+    it('should return disabled thinking when no reasoning effort', async () => {
+      const { isReasoningModel, isSupportedThinkingTokenClaudeModel } = await import('@renderer/config/models')
+
+      vi.mocked(isReasoningModel).mockReturnValue(true)
+      vi.mocked(isSupportedThinkingTokenClaudeModel).mockReturnValue(false)
+
+      const model: Model = {
+        id: 'claude-3-7-sonnet',
+        name: 'Claude 3.7 Sonnet',
+        provider: SystemProviderIds.anthropic
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {}
+      } as Assistant
+
+      const result = getAnthropicReasoningParams(assistant, model)
+      expect(result).toEqual({
+        thinking: {
+          type: 'disabled'
+        }
+      })
+    })
+
+    it('should return enabled thinking with budget for Claude models', async () => {
+      const { isReasoningModel, isSupportedThinkingTokenClaudeModel } = await import('@renderer/config/models')
+
+      vi.mocked(isReasoningModel).mockReturnValue(true)
+      vi.mocked(isSupportedThinkingTokenClaudeModel).mockReturnValue(true)
+
+      const model: Model = {
+        id: 'claude-3-7-sonnet',
+        name: 'Claude 3.7 Sonnet',
+        provider: SystemProviderIds.anthropic
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {
+          reasoning_effort: 'medium',
+          maxTokens: 4096
+        }
+      } as Assistant
+
+      const result = getAnthropicReasoningParams(assistant, model)
+      expect(result).toEqual({
+        thinking: {
+          type: 'enabled',
+          budgetTokens: 2048
+        }
+      })
+    })
+  })
+
+  describe('getGeminiReasoningParams', () => {
+    it('should return empty for non-reasoning model', async () => {
+      const { isReasoningModel } = await import('@renderer/config/models')
+
+      vi.mocked(isReasoningModel).mockReturnValue(false)
+
+      const model: Model = {
+        id: 'gemini-2.0-flash',
+        name: 'Gemini 2.0 Flash',
+        provider: SystemProviderIds.gemini
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {}
+      } as Assistant
+
+      const result = getGeminiReasoningParams(assistant, model)
+      expect(result).toEqual({})
+    })
+
+    it('should disable thinking for Flash models without reasoning effort', async () => {
+      const { isReasoningModel, isSupportedThinkingTokenGeminiModel } = await import('@renderer/config/models')
+
+      vi.mocked(isReasoningModel).mockReturnValue(true)
+      vi.mocked(isSupportedThinkingTokenGeminiModel).mockReturnValue(true)
+
+      const model: Model = {
+        id: 'gemini-2.5-flash',
+        name: 'Gemini 2.5 Flash',
+        provider: SystemProviderIds.gemini
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {}
+      } as Assistant
+
+      const result = getGeminiReasoningParams(assistant, model)
+      expect(result).toEqual({
+        thinkingConfig: {
+          includeThoughts: false,
+          thinkingBudget: 0
+        }
+      })
+    })
+
+    it('should enable thinking with budget for reasoning effort', async () => {
+      const { isReasoningModel, isSupportedThinkingTokenGeminiModel } = await import('@renderer/config/models')
+
+      vi.mocked(isReasoningModel).mockReturnValue(true)
+      vi.mocked(isSupportedThinkingTokenGeminiModel).mockReturnValue(true)
+
+      const model: Model = {
+        id: 'gemini-2.5-pro',
+        name: 'Gemini 2.5 Pro',
+        provider: SystemProviderIds.gemini
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {
+          reasoning_effort: 'medium'
+        }
+      } as Assistant
+
+      const result = getGeminiReasoningParams(assistant, model)
+      expect(result).toEqual({
+        thinkingConfig: {
+          thinkingBudget: 16448,
+          includeThoughts: true
+        }
+      })
+    })
+
+    it('should enable thinking without budget for auto effort ratio > 1', async () => {
+      const { isReasoningModel, isSupportedThinkingTokenGeminiModel } = await import('@renderer/config/models')
+
+      vi.mocked(isReasoningModel).mockReturnValue(true)
+      vi.mocked(isSupportedThinkingTokenGeminiModel).mockReturnValue(true)
+
+      const model: Model = {
+        id: 'gemini-2.5-pro',
+        name: 'Gemini 2.5 Pro',
+        provider: SystemProviderIds.gemini
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {
+          reasoning_effort: 'auto'
+        }
+      } as Assistant
+
+      const result = getGeminiReasoningParams(assistant, model)
+      expect(result).toEqual({
+        thinkingConfig: {
+          includeThoughts: true
+        }
+      })
+    })
+  })
+
+  describe('getXAIReasoningParams', () => {
+    it('should return empty for non-Grok model', async () => {
+      const { isSupportedReasoningEffortGrokModel } = await import('@renderer/config/models')
+
+      vi.mocked(isSupportedReasoningEffortGrokModel).mockReturnValue(false)
+
+      const model: Model = {
+        id: 'other-model',
+        name: 'Other Model',
+        provider: SystemProviderIds.grok
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {}
+      } as Assistant
+
+      const result = getXAIReasoningParams(assistant, model)
+      expect(result).toEqual({})
+    })
+
+    it('should return empty when no reasoning effort', async () => {
+      const { isSupportedReasoningEffortGrokModel } = await import('@renderer/config/models')
+
+      vi.mocked(isSupportedReasoningEffortGrokModel).mockReturnValue(true)
+
+      const model: Model = {
+        id: 'grok-2',
+        name: 'Grok 2',
+        provider: SystemProviderIds.grok
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {}
+      } as Assistant
+
+      const result = getXAIReasoningParams(assistant, model)
+      expect(result).toEqual({})
+    })
+
+    it('should return reasoning effort for Grok models', async () => {
+      const { isSupportedReasoningEffortGrokModel } = await import('@renderer/config/models')
+
+      vi.mocked(isSupportedReasoningEffortGrokModel).mockReturnValue(true)
+
+      const model: Model = {
+        id: 'grok-3',
+        name: 'Grok 3',
+        provider: SystemProviderIds.grok
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {
+          reasoning_effort: 'high'
+        }
+      } as Assistant
+
+      const result = getXAIReasoningParams(assistant, model)
+      expect(result).toHaveProperty('reasoningEffort')
+      expect(result.reasoningEffort).toBe('high')
+    })
+  })
+
+  describe('getBedrockReasoningParams', () => {
+    it('should return empty for non-reasoning model', async () => {
+      const model: Model = {
+        id: 'other-model',
+        name: 'Other Model',
+        provider: 'bedrock'
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {}
+      } as Assistant
+
+      const result = getBedrockReasoningParams(assistant, model)
+      expect(result).toEqual({})
+    })
+
+    it('should return empty when no reasoning effort', async () => {
+      const model: Model = {
+        id: 'claude-3-7-sonnet',
+        name: 'Claude 3.7 Sonnet',
+        provider: 'bedrock'
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {}
+      } as Assistant
+
+      const result = getBedrockReasoningParams(assistant, model)
+      expect(result).toEqual({})
+    })
+
+    it('should return reasoning config for Claude models on Bedrock', async () => {
+      const { isReasoningModel, isSupportedThinkingTokenClaudeModel } = await import('@renderer/config/models')
+
+      vi.mocked(isReasoningModel).mockReturnValue(true)
+      vi.mocked(isSupportedThinkingTokenClaudeModel).mockReturnValue(true)
+
+      const model: Model = {
+        id: 'claude-3-7-sonnet',
+        name: 'Claude 3.7 Sonnet',
+        provider: 'bedrock'
+      } as Model
+
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {
+          reasoning_effort: 'medium',
+          maxTokens: 4096
+        }
+      } as Assistant
+
+      const result = getBedrockReasoningParams(assistant, model)
+      expect(result).toEqual({
+        reasoningConfig: {
+          type: 'enabled',
+          budgetTokens: 2048
+        }
+      })
+    })
+  })
+
+  describe('getCustomParameters', () => {
+    it('should return empty object when no custom parameters', async () => {
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {}
+      } as Assistant
+
+      const result = getCustomParameters(assistant)
+      expect(result).toEqual({})
+    })
+
+    it('should return custom parameters as key-value pairs', async () => {
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {
+          customParameters: [
+            { name: 'param1', value: 'value1', type: 'string' },
+            { name: 'param2', value: 123, type: 'number' }
+          ]
+        }
+      } as Assistant
+
+      const result = getCustomParameters(assistant)
+      expect(result).toEqual({
+        param1: 'value1',
+        param2: 123
+      })
+    })
+
+    it('should parse JSON type parameters', async () => {
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {
+          customParameters: [{ name: 'config', value: '{"key": "value"}', type: 'json' }]
+        }
+      } as Assistant
+
+      const result = getCustomParameters(assistant)
+      expect(result).toEqual({
+        config: { key: 'value' }
+      })
+    })
+
+    it('should handle invalid JSON gracefully', async () => {
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {
+          customParameters: [{ name: 'invalid', value: '{invalid json', type: 'json' }]
+        }
+      } as Assistant
+
+      const result = getCustomParameters(assistant)
+      expect(result).toEqual({
+        invalid: '{invalid json'
+      })
+    })
+
+    it('should handle undefined JSON value', async () => {
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {
+          customParameters: [{ name: 'undef', value: 'undefined', type: 'json' }]
+        }
+      } as Assistant
+
+      const result = getCustomParameters(assistant)
+      expect(result).toEqual({
+        undef: undefined
+      })
+    })
+
+    it('should skip parameters with empty names', async () => {
+      const assistant: Assistant = {
+        id: 'test',
+        name: 'Test',
+        settings: {
+          customParameters: [
+            { name: '', value: 'value1', type: 'string' },
+            { name: '  ', value: 'value2', type: 'string' },
+            { name: 'valid', value: 'value3', type: 'string' }
+          ]
+        }
+      } as Assistant
+
+      const result = getCustomParameters(assistant)
+      expect(result).toEqual({
+        valid: 'value3'
+      })
+    })
+  })
+})
@@ -0,0 +1,384 @@
+/**
+ * websearch.ts Unit Tests
+ * Tests for web search parameters generation utilities
+ */
+
+import type { CherryWebSearchConfig } from '@renderer/store/websearch'
+import type { Model } from '@renderer/types'
+import { describe, expect, it, vi } from 'vitest'
+
+import { buildProviderBuiltinWebSearchConfig, getWebSearchParams } from '../websearch'
+
+// Mock dependencies
+vi.mock('@renderer/config/models', () => ({
+  isOpenAIWebSearchChatCompletionOnlyModel: vi.fn((model) => model?.id?.includes('o1-pro') ?? false),
+  isOpenAIDeepResearchModel: vi.fn((model) => model?.id?.includes('o3-mini') ?? false)
+}))
+
+vi.mock('@renderer/utils/blacklistMatchPattern', () => ({
+  mapRegexToPatterns: vi.fn((patterns) => patterns || [])
+}))
+
+describe('websearch utils', () => {
+  describe('getWebSearchParams', () => {
+    it('should return enhancement params for hunyuan provider', () => {
+      const model: Model = {
+        id: 'hunyuan-model',
+        name: 'Hunyuan Model',
+        provider: 'hunyuan'
+      } as Model
+
+      const result = getWebSearchParams(model)
+
+      expect(result).toEqual({
+        enable_enhancement: true,
+        citation: true,
+        search_info: true
+      })
+    })
+
+    it('should return search params for dashscope provider', () => {
+      const model: Model = {
+        id: 'qwen-model',
+        name: 'Qwen Model',
+        provider: 'dashscope'
+      } as Model
+
+      const result = getWebSearchParams(model)
+
+      expect(result).toEqual({
+        enable_search: true,
+        search_options: {
+          forced_search: true
+        }
+      })
+    })
+
+    it('should return web_search_options for OpenAI web search models', () => {
+      const model: Model = {
+        id: 'o1-pro',
+        name: 'O1 Pro',
+        provider: 'openai'
+      } as Model
+
+      const result = getWebSearchParams(model)
+
+      expect(result).toEqual({
+        web_search_options: {}
+      })
+    })
+
+    it('should return empty object for other providers', () => {
+      const model: Model = {
+        id: 'gpt-4',
+        name: 'GPT-4',
+        provider: 'openai'
+      } as Model
+
+      const result = getWebSearchParams(model)
+
+      expect(result).toEqual({})
+    })
+
+    it('should return empty object for custom provider', () => {
+      const model: Model = {
+        id: 'custom-model',
+        name: 'Custom Model',
+        provider: 'custom-provider'
+      } as Model
+
+      const result = getWebSearchParams(model)
+
+      expect(result).toEqual({})
+    })
+  })
+
+  describe('buildProviderBuiltinWebSearchConfig', () => {
+    const defaultWebSearchConfig: CherryWebSearchConfig = {
+      searchWithTime: true,
+      maxResults: 50,
+      excludeDomains: []
+    }
+
+    describe('openai provider', () => {
+      it('should return low search context size for low maxResults', () => {
+        const config: CherryWebSearchConfig = {
+          searchWithTime: true,
+          maxResults: 20,
+          excludeDomains: []
+        }
+
+        const result = buildProviderBuiltinWebSearchConfig('openai', config)
+
+        expect(result).toEqual({
+          openai: {
+            searchContextSize: 'low'
+          }
+        })
+      })
+
+      it('should return medium search context size for medium maxResults', () => {
+        const config: CherryWebSearchConfig = {
+          searchWithTime: true,
+          maxResults: 50,
+          excludeDomains: []
+        }
+
+        const result = buildProviderBuiltinWebSearchConfig('openai', config)
+
+        expect(result).toEqual({
+          openai: {
+            searchContextSize: 'medium'
+          }
+        })
+      })
+
+      it('should return high search context size for high maxResults', () => {
+        const config: CherryWebSearchConfig = {
+          searchWithTime: true,
+          maxResults: 80,
+          excludeDomains: []
+        }
+
+        const result = buildProviderBuiltinWebSearchConfig('openai', config)
+
+        expect(result).toEqual({
+          openai: {
+            searchContextSize: 'high'
+          }
+        })
+      })
+
+      it('should use medium for deep research models regardless of maxResults', () => {
+        const config: CherryWebSearchConfig = {
+          searchWithTime: true,
+          maxResults: 100,
+          excludeDomains: []
+        }
+
+        const model: Model = {
+          id: 'o3-mini',
+          name: 'O3 Mini',
+          provider: 'openai'
+        } as Model
+
+        const result = buildProviderBuiltinWebSearchConfig('openai', config, model)
+
+        expect(result).toEqual({
+          openai: {
+            searchContextSize: 'medium'
+          }
+        })
+      })
+    })
+
+    describe('openai-chat provider', () => {
+      it('should return correct search context size', () => {
+        const config: CherryWebSearchConfig = {
+          searchWithTime: true,
+          maxResults: 50,
+          excludeDomains: []
+        }
+
+        const result = buildProviderBuiltinWebSearchConfig('openai-chat', config)
+
+        expect(result).toEqual({
+          'openai-chat': {
+            searchContextSize: 'medium'
+          }
+        })
+      })
+
+      it('should handle deep research models', () => {
+        const config: CherryWebSearchConfig = {
+          searchWithTime: true,
+          maxResults: 100,
+          excludeDomains: []
+        }
+
+        const model: Model = {
+          id: 'o3-mini',
+          name: 'O3 Mini',
+          provider: 'openai'
+        } as Model
+
+        const result = buildProviderBuiltinWebSearchConfig('openai-chat', config, model)
+
+        expect(result).toEqual({
+          'openai-chat': {
+            searchContextSize: 'medium'
+          }
+        })
+      })
+    })
+
+    describe('anthropic provider', () => {
+      it('should return anthropic search options with maxUses', () => {
+        const result = buildProviderBuiltinWebSearchConfig('anthropic', defaultWebSearchConfig)
+
+        expect(result).toEqual({
+          anthropic: {
+            maxUses: 50,
+            blockedDomains: undefined
+          }
+        })
+      })
+
+      it('should include blockedDomains when excludeDomains provided', () => {
+        const config: CherryWebSearchConfig = {
+          searchWithTime: true,
+          maxResults: 30,
+          excludeDomains: ['example.com', 'test.com']
+        }
+
+        const result = buildProviderBuiltinWebSearchConfig('anthropic', config)
+
+        expect(result).toEqual({
+          anthropic: {
+            maxUses: 30,
+            blockedDomains: ['example.com', 'test.com']
+          }
+        })
+      })
+
+      it('should not include blockedDomains when empty', () => {
+        const result = buildProviderBuiltinWebSearchConfig('anthropic', defaultWebSearchConfig)
+
+        expect(result).toEqual({
+          anthropic: {
+            maxUses: 50,
+            blockedDomains: undefined
+          }
+        })
+      })
+    })
+
+    describe('xai provider', () => {
+      it('should return xai search options', () => {
+        const result = buildProviderBuiltinWebSearchConfig('xai', defaultWebSearchConfig)
+
+        expect(result).toEqual({
+          xai: {
+            maxSearchResults: 50,
+            returnCitations: true,
+            sources: [{ type: 'web', excludedWebsites: [] }, { type: 'news' }, { type: 'x' }],
+            mode: 'on'
+          }
+        })
+      })
+
+      it('should limit excluded websites to 5', () => {
+        const config: CherryWebSearchConfig = {
+          searchWithTime: true,
+          maxResults: 40,
+          excludeDomains: ['site1.com', 'site2.com', 'site3.com', 'site4.com', 'site5.com', 'site6.com', 'site7.com']
+        }
+
+        const result = buildProviderBuiltinWebSearchConfig('xai', config)
+
+        expect(result?.xai?.sources).toBeDefined()
+        const webSource = result?.xai?.sources?.[0]
+        if (webSource && webSource.type === 'web') {
+          expect(webSource.excludedWebsites).toHaveLength(5)
+        }
+      })
+
+      it('should include all sources types', () => {
+        const result = buildProviderBuiltinWebSearchConfig('xai', defaultWebSearchConfig)
+
+        expect(result?.xai?.sources).toHaveLength(3)
+        expect(result?.xai?.sources?.[0].type).toBe('web')
+        expect(result?.xai?.sources?.[1].type).toBe('news')
+        expect(result?.xai?.sources?.[2].type).toBe('x')
+      })
+    })
+
+    describe('openrouter provider', () => {
+      it('should return openrouter plugins config', () => {
+        const result = buildProviderBuiltinWebSearchConfig('openrouter', defaultWebSearchConfig)
+
+        expect(result).toEqual({
+          openrouter: {
+            plugins: [
+              {
+                id: 'web',
+                max_results: 50
+              }
+            ]
+          }
+        })
+      })
+
+      it('should respect custom maxResults', () => {
+        const config: CherryWebSearchConfig = {
+          searchWithTime: true,
+          maxResults: 75,
+          excludeDomains: []
+        }
+
+        const result = buildProviderBuiltinWebSearchConfig('openrouter', config)
+
+        expect(result).toEqual({
+          openrouter: {
+            plugins: [
+              {
+                id: 'web',
+                max_results: 75
+              }
+            ]
+          }
+        })
+      })
+    })
+
+    describe('unsupported provider', () => {
+      it('should return empty object for unsupported provider', () => {
+        const result = buildProviderBuiltinWebSearchConfig('unsupported' as any, defaultWebSearchConfig)
+
+        expect(result).toEqual({})
+      })
+
+      it('should return empty object for google provider', () => {
+        const result = buildProviderBuiltinWebSearchConfig('google', defaultWebSearchConfig)
+
+        expect(result).toEqual({})
+      })
+    })
+
+    describe('edge cases', () => {
+      it('should handle maxResults at boundary values', () => {
+        // Test boundary at 33 (low/medium)
+        const config33: CherryWebSearchConfig = { searchWithTime: true, maxResults: 33, excludeDomains: [] }
+        const result33 = buildProviderBuiltinWebSearchConfig('openai', config33)
+        expect(result33?.openai?.searchContextSize).toBe('low')
+
+        // Test boundary at 34 (medium)
+        const config34: CherryWebSearchConfig = { searchWithTime: true, maxResults: 34, excludeDomains: [] }
+        const result34 = buildProviderBuiltinWebSearchConfig('openai', config34)
+        expect(result34?.openai?.searchContextSize).toBe('medium')
+
+        // Test boundary at 66 (medium)
+        const config66: CherryWebSearchConfig = { searchWithTime: true, maxResults: 66, excludeDomains: [] }
+        const result66 = buildProviderBuiltinWebSearchConfig('openai', config66)
+        expect(result66?.openai?.searchContextSize).toBe('medium')
+
+        // Test boundary at 67 (high)
+        const config67: CherryWebSearchConfig = { searchWithTime: true, maxResults: 67, excludeDomains: [] }
+        const result67 = buildProviderBuiltinWebSearchConfig('openai', config67)
+        expect(result67?.openai?.searchContextSize).toBe('high')
+      })
+
+      it('should handle zero maxResults', () => {
+        const config: CherryWebSearchConfig = { searchWithTime: true, maxResults: 0, excludeDomains: [] }
+        const result = buildProviderBuiltinWebSearchConfig('openai', config)
+        expect(result?.openai?.searchContextSize).toBe('low')
+      })
+
+      it('should handle very large maxResults', () => {
+        const config: CherryWebSearchConfig = { searchWithTime: true, maxResults: 1000, excludeDomains: [] }
+        const result = buildProviderBuiltinWebSearchConfig('openai', config)
+        expect(result?.openai?.searchContextSize).toBe('high')
+      })
+    })
+  })
+})
@@ -1,16 +1,38 @@
+import type { BedrockProviderOptions } from '@ai-sdk/amazon-bedrock'
+import type { AnthropicProviderOptions } from '@ai-sdk/anthropic'
+import type { GoogleGenerativeAIProviderOptions } from '@ai-sdk/google'
+import type { OpenAIResponsesProviderOptions } from '@ai-sdk/openai'
+import type { XaiProviderOptions } from '@ai-sdk/xai'
 import { baseProviderIdSchema, customProviderIdSchema } from '@cherrystudio/ai-core/provider'
-import { isOpenAIModel, isQwenMTModel, isSupportFlexServiceTierModel } from '@renderer/config/models'
-import { isSupportServiceTierProvider } from '@renderer/config/providers'
-import { mapLanguageToQwenMTModel } from '@renderer/config/translate'
-import type { Assistant, Model, Provider } from '@renderer/types'
+import { loggerService } from '@logger'
 import {
+  getModelSupportedVerbosity,
+  isOpenAIModel,
+  isQwenMTModel,
+  isSupportFlexServiceTierModel,
+  isSupportVerbosityModel
+} from '@renderer/config/models'
+import { mapLanguageToQwenMTModel } from '@renderer/config/translate'
+import { getStoreSetting } from '@renderer/hooks/useSettings'
+import {
+  type Assistant,
+  type GroqServiceTier,
  GroqServiceTiers,
+  type GroqSystemProvider,
  isGroqServiceTier,
+  isGroqSystemProvider,
  isOpenAIServiceTier,
  isTranslateAssistant,
+  type Model,
+  type NotGroqProvider,
+  type OpenAIServiceTier,
  OpenAIServiceTiers,
-  SystemProviderIds
+  type Provider,
+  type ServiceTier
 } from '@renderer/types'
+import type { OpenAIVerbosity } from '@renderer/types/aiCoreTypes'
+import { isSupportServiceTierProvider } from '@renderer/utils/provider'
+import type { JSONValue } from 'ai'
 import { t } from 'i18next'

 import { getAiSdkProviderId } from '../provider/factory'
@@ -26,8 +48,33 @@ import {
 } from './reasoning'
 import { getWebSearchParams } from './websearch'

-// copy from BaseApiClient.ts
-const getServiceTier = (model: Model, provider: Provider) => {
+const logger = loggerService.withContext('aiCore.utils.options')
+
+function toOpenAIServiceTier(model: Model, serviceTier: ServiceTier): OpenAIServiceTier {
+  if (
+    !isOpenAIServiceTier(serviceTier) ||
+    (serviceTier === OpenAIServiceTiers.flex && !isSupportFlexServiceTierModel(model))
+  ) {
+    return undefined
+  } else {
+    return serviceTier
+  }
+}
+
+function toGroqServiceTier(model: Model, serviceTier: ServiceTier): GroqServiceTier {
+  if (
+    !isGroqServiceTier(serviceTier) ||
+    (serviceTier === GroqServiceTiers.flex && !isSupportFlexServiceTierModel(model))
+  ) {
+    return undefined
+  } else {
+    return serviceTier
+  }
+}
+
+function getServiceTier<T extends GroqSystemProvider>(model: Model, provider: T): GroqServiceTier
+function getServiceTier<T extends NotGroqProvider>(model: Model, provider: T): OpenAIServiceTier
+function getServiceTier<T extends Provider>(model: Model, provider: T): OpenAIServiceTier | GroqServiceTier {
  const serviceTierSetting = provider.serviceTier

  if (!isSupportServiceTierProvider(provider) || !isOpenAIModel(model) || !serviceTierSetting) {
@@ -35,24 +82,17 @@ const getServiceTier = (model: Model, provider: Provider) => {
  }

  // 处理不同供应商需要 fallback 到默认值的情况
-  if (provider.id === SystemProviderIds.groq) {
-    if (
-      !isGroqServiceTier(serviceTierSetting) ||
-      (serviceTierSetting === GroqServiceTiers.flex && !isSupportFlexServiceTierModel(model))
-    ) {
-      return undefined
-    }
+  if (isGroqSystemProvider(provider)) {
+    return toGroqServiceTier(model, serviceTierSetting)
  } else {
    // 其他 OpenAI 供应商，假设他们的服务层级设置和 OpenAI 完全相同
-    if (
-      !isOpenAIServiceTier(serviceTierSetting) ||
-      (serviceTierSetting === OpenAIServiceTiers.flex && !isSupportFlexServiceTierModel(model))
-    ) {
-      return undefined
-    }
+    return toOpenAIServiceTier(model, serviceTierSetting)
  }
+}

-  return serviceTierSetting
+function getVerbosity(): OpenAIVerbosity {
+  const openAI = getStoreSetting('openAI')
+  return openAI.verbosity
 }

 /**
@@ -69,12 +109,13 @@ export function buildProviderOptions(
    enableWebSearch: boolean
    enableGenerateImage: boolean
  }
-): Record<string, any> {
+): Record<string, Record<string, JSONValue>> {
+  logger.debug('buildProviderOptions', { assistant, model, actualProvider, capabilities })
  const rawProviderId = getAiSdkProviderId(actualProvider)
  // 构建 provider 特定的选项
  let providerSpecificOptions: Record<string, any> = {}
-  const serviceTierSetting = getServiceTier(model, actualProvider)
-  providerSpecificOptions.serviceTier = serviceTierSetting
+  const serviceTier = getServiceTier(model, actualProvider)
+  const textVerbosity = getVerbosity()
  // 根据 provider 类型分离构建逻辑
  const { data: baseProviderId, success } = baseProviderIdSchema.safeParse(rawProviderId)
  if (success) {
@@ -84,14 +125,16 @@ export function buildProviderOptions(
      case 'openai-chat':
      case 'azure':
      case 'azure-responses':
-        providerSpecificOptions = {
-          ...buildOpenAIProviderOptions(assistant, model, capabilities),
-          serviceTier: serviceTierSetting
+        {
+          const options: OpenAIResponsesProviderOptions = buildOpenAIProviderOptions(
+            assistant,
+            model,
+            capabilities,
+            serviceTier
+          )
+          providerSpecificOptions = options
        }
        break
-      case 'huggingface':
-        providerSpecificOptions = buildOpenAIProviderOptions(assistant, model, capabilities)
-        break
      case 'anthropic':
        providerSpecificOptions = buildAnthropicProviderOptions(assistant, model, capabilities)
        break
@@ -109,12 +152,19 @@ export function buildProviderOptions(
        // 对于其他 provider，使用通用的构建逻辑
        providerSpecificOptions = {
          ...buildGenericProviderOptions(assistant, model, capabilities),
-          serviceTier: serviceTierSetting
+          serviceTier,
+          textVerbosity
        }
        break
      }
      case 'cherryin':
-        providerSpecificOptions = buildCherryInProviderOptions(assistant, model, capabilities, actualProvider)
+        providerSpecificOptions = buildCherryInProviderOptions(
+          assistant,
+          model,
+          capabilities,
+          actualProvider,
+          serviceTier
+        )
        break
      default:
        throw new Error(`Unsupported base provider ${baseProviderId}`)
@@ -128,17 +178,22 @@ export function buildProviderOptions(
        case 'google-vertex':
          providerSpecificOptions = buildGeminiProviderOptions(assistant, model, capabilities)
          break
+        case 'azure-anthropic':
        case 'google-vertex-anthropic':
          providerSpecificOptions = buildAnthropicProviderOptions(assistant, model, capabilities)
          break
        case 'bedrock':
          providerSpecificOptions = buildBedrockProviderOptions(assistant, model, capabilities)
          break
+        case 'huggingface':
+          providerSpecificOptions = buildOpenAIProviderOptions(assistant, model, capabilities, serviceTier)
+          break
        default:
          // 对于其他 provider，使用通用的构建逻辑
          providerSpecificOptions = {
            ...buildGenericProviderOptions(assistant, model, capabilities),
-            serviceTier: serviceTierSetting
+            serviceTier,
+            textVerbosity
          }
      }
    } else {
@@ -152,13 +207,18 @@ export function buildProviderOptions(
    ...getCustomParameters(assistant)
  }

-  const rawProviderKey =
+  let rawProviderKey =
    {
      'google-vertex': 'google',
      'google-vertex-anthropic': 'anthropic',
+      'azure-anthropic': 'anthropic',
      'ai-gateway': 'gateway'
    }[rawProviderId] || rawProviderId

+  if (rawProviderKey === 'cherryin') {
+    rawProviderKey = { gemini: 'google' }[actualProvider.type] || actualProvider.type
+  }
+
  // 返回 AI Core SDK 要求的格式：{ 'providerId': providerOptions }
  return {
    [rawProviderKey]: providerSpecificOptions
@@ -175,10 +235,11 @@ function buildOpenAIProviderOptions(
    enableReasoning: boolean
    enableWebSearch: boolean
    enableGenerateImage: boolean
-  }
-): Record<string, any> {
+  },
+  serviceTier: OpenAIServiceTier
+): OpenAIResponsesProviderOptions {
  const { enableReasoning } = capabilities
-  let providerOptions: Record<string, any> = {}
+  let providerOptions: OpenAIResponsesProviderOptions = {}
  // OpenAI 推理参数
  if (enableReasoning) {
    const reasoningParams = getOpenAIReasoningParams(assistant, model)
@@ -187,6 +248,28 @@ function buildOpenAIProviderOptions(
      ...reasoningParams
    }
  }
+
+  if (isSupportVerbosityModel(model)) {
+    const openAI = getStoreSetting<'openAI'>('openAI')
+    const userVerbosity = openAI?.verbosity
+
+    if (userVerbosity && ['low', 'medium', 'high'].includes(userVerbosity)) {
+      const supportedVerbosity = getModelSupportedVerbosity(model)
+      // Use user's verbosity if supported, otherwise use the first supported option
+      const verbosity = supportedVerbosity.includes(userVerbosity) ? userVerbosity : supportedVerbosity[0]
+
+      providerOptions = {
+        ...providerOptions,
+        textVerbosity: verbosity
+      }
+    }
+  }
+
+  providerOptions = {
+    ...providerOptions,
+    serviceTier
+  }
+
  return providerOptions
 }

@@ -201,9 +284,9 @@ function buildAnthropicProviderOptions(
    enableWebSearch: boolean
    enableGenerateImage: boolean
  }
-): Record<string, any> {
+): AnthropicProviderOptions {
  const { enableReasoning } = capabilities
-  let providerOptions: Record<string, any> = {}
+  let providerOptions: AnthropicProviderOptions = {}

  // Anthropic 推理参数
  if (enableReasoning) {
@@ -228,9 +311,9 @@ function buildGeminiProviderOptions(
    enableWebSearch: boolean
    enableGenerateImage: boolean
  }
-): Record<string, any> {
+): GoogleGenerativeAIProviderOptions {
  const { enableReasoning, enableGenerateImage } = capabilities
-  let providerOptions: Record<string, any> = {}
+  let providerOptions: GoogleGenerativeAIProviderOptions = {}

  // Gemini 推理参数
  if (enableReasoning) {
@@ -259,7 +342,7 @@ function buildXAIProviderOptions(
    enableWebSearch: boolean
    enableGenerateImage: boolean
  }
-): Record<string, any> {
+): XaiProviderOptions {
  const { enableReasoning } = capabilities
  let providerOptions: Record<string, any> = {}

@@ -282,16 +365,12 @@ function buildCherryInProviderOptions(
    enableWebSearch: boolean
    enableGenerateImage: boolean
  },
-  actualProvider: Provider
-): Record<string, any> {
-  const serviceTierSetting = getServiceTier(model, actualProvider)
-
+  actualProvider: Provider,
+  serviceTier: OpenAIServiceTier
+): OpenAIResponsesProviderOptions | AnthropicProviderOptions | GoogleGenerativeAIProviderOptions {
  switch (actualProvider.type) {
    case 'openai':
-      return {
-        ...buildOpenAIProviderOptions(assistant, model, capabilities),
-        serviceTier: serviceTierSetting
-      }
+      return buildOpenAIProviderOptions(assistant, model, capabilities, serviceTier)

    case 'anthropic':
      return buildAnthropicProviderOptions(assistant, model, capabilities)
@@ -313,9 +392,9 @@ function buildBedrockProviderOptions(
    enableWebSearch: boolean
    enableGenerateImage: boolean
  }
-): Record<string, any> {
+): BedrockProviderOptions {
  const { enableReasoning } = capabilities
-  let providerOptions: Record<string, any> = {}
+  let providerOptions: BedrockProviderOptions = {}

  if (enableReasoning) {
    const reasoningParams = getBedrockReasoningParams(assistant, model)
@@ -1,3 +1,8 @@
+import type { BedrockProviderOptions } from '@ai-sdk/amazon-bedrock'
+import type { AnthropicProviderOptions } from '@ai-sdk/anthropic'
+import type { GoogleGenerativeAIProviderOptions } from '@ai-sdk/google'
+import type { OpenAIResponsesProviderOptions } from '@ai-sdk/openai'
+import type { XaiProviderOptions } from '@ai-sdk/xai'
 import { loggerService } from '@logger'
 import { DEFAULT_MAX_TOKENS } from '@renderer/config/constant'
 import {
@@ -7,6 +12,8 @@ import {
  isDeepSeekHybridInferenceModel,
  isDoubaoSeedAfter251015,
  isDoubaoThinkingAutoModel,
+  isGemini3Model,
+  isGPT51SeriesModel,
  isGrok4FastReasoningModel,
  isGrokReasoningModel,
  isOpenAIDeepResearchModel,
@@ -27,13 +34,13 @@ import {
  isSupportedThinkingTokenZhipuModel,
  MODEL_SUPPORTED_REASONING_EFFORT
 } from '@renderer/config/models'
-import { isSupportEnableThinkingProvider } from '@renderer/config/providers'
 import { getStoreSetting } from '@renderer/hooks/useSettings'
 import { getAssistantSettings, getProviderByModel } from '@renderer/services/AssistantService'
-import type { SettingsState } from '@renderer/store/settings'
-import type { Assistant, Model } from '@renderer/types'
+import type { Assistant, Model, ReasoningEffortOption } from '@renderer/types'
 import { EFFORT_RATIO, isSystemProvider, SystemProviderIds } from '@renderer/types'
+import type { OpenAISummaryText } from '@renderer/types/aiCoreTypes'
 import type { ReasoningEffortOptionalParams } from '@renderer/types/sdk'
+import { isSupportEnableThinkingProvider } from '@renderer/utils/provider'
 import { toInteger } from 'lodash'

 const logger = loggerService.withContext('reasoning')
@@ -56,13 +63,20 @@ export function getReasoningEffort(assistant: Assistant, model: Model): Reasonin
  }
  const reasoningEffort = assistant?.settings?.reasoning_effort

-  if (!reasoningEffort) {
+  // Handle undefined and 'none' reasoningEffort.
+  // TODO: They should be separated.
+  if (!reasoningEffort || reasoningEffort === 'none') {
    // openrouter: use reasoning
    if (model.provider === SystemProviderIds.openrouter) {
      // Don't disable reasoning for Gemini models that support thinking tokens
      if (isSupportedThinkingTokenGeminiModel(model) && !GEMINI_FLASH_MODEL_REGEX.test(model.id)) {
        return {}
      }
+      // 'none' is not an available value for effort for now.
+      // I think they should resolve this issue soon, so I'll just go ahead and use this value.
+      if (isGPT51SeriesModel(model) && reasoningEffort === 'none') {
+        return { reasoning: { effort: 'none' } }
+      }
      // Don't disable reasoning for models that require it
      if (
        isGrokReasoningModel(model) ||
@@ -117,6 +131,13 @@ export function getReasoningEffort(assistant: Assistant, model: Model): Reasonin
      return { thinking: { type: 'disabled' } }
    }

+    // Specially for GPT-5.1. Suppose this is a OpenAI Compatible provider
+    if (isGPT51SeriesModel(model)) {
+      return {
+        reasoningEffort: 'none'
+      }
+    }
+
    return {}
  }

@@ -259,6 +280,12 @@ export function getReasoningEffort(assistant: Assistant, model: Model): Reasonin

  // gemini series, openai compatible api
  if (isSupportedThinkingTokenGeminiModel(model)) {
+    // https://ai.google.dev/gemini-api/docs/gemini-3?thinking=high#openai_compatibility
+    if (isGemini3Model(model)) {
+      return {
+        reasoning_effort: reasoningEffort
+      }
+    }
    if (reasoningEffort === 'auto') {
      return {
        extra_body: {
@@ -322,10 +349,14 @@ export function getReasoningEffort(assistant: Assistant, model: Model): Reasonin
 }

 /**
- * 获取 OpenAI 推理参数
- * 从 OpenAIResponseAPIClient 和 OpenAIAPIClient 中提取的逻辑
+ * Get OpenAI reasoning parameters
+ * Extracted from OpenAIResponseAPIClient and OpenAIAPIClient logic
+ * For official OpenAI provider only
 */
-export function getOpenAIReasoningParams(assistant: Assistant, model: Model): Record<string, any> {
+export function getOpenAIReasoningParams(
+  assistant: Assistant,
+  model: Model
+): Pick<OpenAIResponsesProviderOptions, 'reasoningEffort' | 'reasoningSummary'> {
  if (!isReasoningModel(model)) {
    return {}
  }
@@ -336,6 +367,10 @@ export function getOpenAIReasoningParams(assistant: Assistant, model: Model): Re
    return {}
  }

+  if (isOpenAIDeepResearchModel(model) || reasoningEffort === 'auto') {
+    reasoningEffort = 'medium'
+  }
+
  // 非OpenAI模型，但是Provider类型是responses/azure openai的情况
  if (!isOpenAIModel(model)) {
    return {
@@ -343,21 +378,17 @@ export function getOpenAIReasoningParams(assistant: Assistant, model: Model): Re
    }
  }

-  const openAI = getStoreSetting('openAI') as SettingsState['openAI']
-  const summaryText = openAI?.summaryText || 'off'
+  const openAI = getStoreSetting('openAI')
+  const summaryText = openAI.summaryText

-  let reasoningSummary: string | undefined = undefined
+  let reasoningSummary: OpenAISummaryText = undefined

-  if (summaryText === 'off' || model.id.includes('o1-pro')) {
+  if (model.id.includes('o1-pro')) {
    reasoningSummary = undefined
  } else {
    reasoningSummary = summaryText
  }

-  if (isOpenAIDeepResearchModel(model)) {
-    reasoningEffort = 'medium'
-  }
-
  // OpenAI 推理参数
  if (isSupportedReasoningEffortOpenAIModel(model)) {
    return {
@@ -369,19 +400,26 @@ export function getOpenAIReasoningParams(assistant: Assistant, model: Model): Re
  return {}
 }

-export function getAnthropicThinkingBudget(assistant: Assistant, model: Model): number {
-  const { maxTokens, reasoning_effort: reasoningEffort } = getAssistantSettings(assistant)
-  if (reasoningEffort === undefined) {
-    return 0
+export function getAnthropicThinkingBudget(
+  maxTokens: number | undefined,
+  reasoningEffort: string | undefined,
+  modelId: string
+): number | undefined {
+  if (reasoningEffort === undefined || reasoningEffort === 'none') {
+    return undefined
  }
  const effortRatio = EFFORT_RATIO[reasoningEffort]

+  const tokenLimit = findTokenLimit(modelId)
+  if (!tokenLimit) {
+    return undefined
+  }
+
  const budgetTokens = Math.max(
    1024,
    Math.floor(
      Math.min(
-        (findTokenLimit(model.id)?.max! - findTokenLimit(model.id)?.min!) * effortRatio +
-          findTokenLimit(model.id)?.min!,
+        (tokenLimit.max - tokenLimit.min) * effortRatio + tokenLimit.min,
        (maxTokens || DEFAULT_MAX_TOKENS) * effortRatio
      )
    )
@@ -393,14 +431,17 @@ export function getAnthropicThinkingBudget(assistant: Assistant, model: Model):
 * 获取 Anthropic 推理参数
 * 从 AnthropicAPIClient 中提取的逻辑
 */
-export function getAnthropicReasoningParams(assistant: Assistant, model: Model): Record<string, any> {
+export function getAnthropicReasoningParams(
+  assistant: Assistant,
+  model: Model
+): Pick<AnthropicProviderOptions, 'thinking'> {
  if (!isReasoningModel(model)) {
    return {}
  }

  const reasoningEffort = assistant?.settings?.reasoning_effort

-  if (reasoningEffort === undefined) {
+  if (reasoningEffort === undefined || reasoningEffort === 'none') {
    return {
      thinking: {
        type: 'disabled'
@@ -410,7 +451,8 @@ export function getAnthropicReasoningParams(assistant: Assistant, model: Model):

  // Claude 推理参数
  if (isSupportedThinkingTokenClaudeModel(model)) {
-    const budgetTokens = getAnthropicThinkingBudget(assistant, model)
+    const { maxTokens } = getAssistantSettings(assistant)
+    const budgetTokens = getAnthropicThinkingBudget(maxTokens, reasoningEffort, model.id)

    return {
      thinking: {
@@ -423,13 +465,31 @@ export function getAnthropicReasoningParams(assistant: Assistant, model: Model):
  return {}
 }

+type GoogelThinkingLevel = NonNullable<GoogleGenerativeAIProviderOptions['thinkingConfig']>['thinkingLevel']
+
+function mapToGeminiThinkingLevel(reasoningEffort: ReasoningEffortOption): GoogelThinkingLevel {
+  switch (reasoningEffort) {
+    case 'low':
+      return 'low'
+    case 'medium':
+      return 'medium'
+    case 'high':
+      return 'high'
+    default:
+      return 'medium'
+  }
+}
+
 /**
 * 获取 Gemini 推理参数
 * 从 GeminiAPIClient 中提取的逻辑
 * 注意：Gemini/GCP 端点所使用的 thinkingBudget 等参数应该按照驼峰命名法传递
 * 而在 Google 官方提供的 OpenAI 兼容端点中则使用蛇形命名法 thinking_budget
 */
-export function getGeminiReasoningParams(assistant: Assistant, model: Model): Record<string, any> {
+export function getGeminiReasoningParams(
+  assistant: Assistant,
+  model: Model
+): Pick<GoogleGenerativeAIProviderOptions, 'thinkingConfig'> {
  if (!isReasoningModel(model)) {
    return {}
  }
@@ -438,7 +498,7 @@ export function getGeminiReasoningParams(assistant: Assistant, model: Model): Re

  // Gemini 推理参数
  if (isSupportedThinkingTokenGeminiModel(model)) {
-    if (reasoningEffort === undefined) {
+    if (reasoningEffort === undefined || reasoningEffort === 'none') {
      return {
        thinkingConfig: {
          includeThoughts: false,
@@ -447,6 +507,15 @@ export function getGeminiReasoningParams(assistant: Assistant, model: Model): Re
      }
    }

+    // https://ai.google.dev/gemini-api/docs/gemini-3?thinking=high#new_api_features_in_gemini_3
+    if (isGemini3Model(model)) {
+      return {
+        thinkingConfig: {
+          thinkingLevel: mapToGeminiThinkingLevel(reasoningEffort)
+        }
+      }
+    }
+
    const effortRatio = EFFORT_RATIO[reasoningEffort]

    if (effortRatio > 1) {
@@ -478,27 +547,35 @@ export function getGeminiReasoningParams(assistant: Assistant, model: Model): Re
 * @param model - The model being used
 * @returns XAI-specific reasoning parameters
 */
-export function getXAIReasoningParams(assistant: Assistant, model: Model): Record<string, any> {
+export function getXAIReasoningParams(assistant: Assistant, model: Model): Pick<XaiProviderOptions, 'reasoningEffort'> {
  if (!isSupportedReasoningEffortGrokModel(model)) {
    return {}
  }

  const { reasoning_effort: reasoningEffort } = getAssistantSettings(assistant)

-  if (!reasoningEffort) {
+  if (!reasoningEffort || reasoningEffort === 'none') {
    return {}
  }

-  // For XAI provider Grok models, use reasoningEffort parameter directly
-  return {
-    reasoningEffort
+  switch (reasoningEffort) {
+    case 'auto':
+    case 'minimal':
+    case 'medium':
+      return { reasoningEffort: 'low' }
+    case 'low':
+    case 'high':
+      return { reasoningEffort }
  }
 }

 /**
 * Get Bedrock reasoning parameters
 */
-export function getBedrockReasoningParams(assistant: Assistant, model: Model): Record<string, any> {
+export function getBedrockReasoningParams(
+  assistant: Assistant,
+  model: Model
+): Pick<BedrockProviderOptions, 'reasoningConfig'> {
  if (!isReasoningModel(model)) {
    return {}
  }
@@ -509,12 +586,21 @@ export function getBedrockReasoningParams(assistant: Assistant, model: Model): R
    return {}
  }

+  if (reasoningEffort === 'none') {
+    return {
+      reasoningConfig: {
+        type: 'disabled'
+      }
+    }
+  }
+
  // Only apply thinking budget for Claude reasoning models
  if (!isSupportedThinkingTokenClaudeModel(model)) {
    return {}
  }

-  const budgetTokens = getAnthropicThinkingBudget(assistant, model)
+  const { maxTokens } = getAssistantSettings(assistant)
+  const budgetTokens = getAnthropicThinkingBudget(maxTokens, reasoningEffort, model.id)
  return {
    reasoningConfig: {
      type: 'enabled',
@@ -47,6 +47,7 @@ export function buildProviderBuiltinWebSearchConfig(
  model?: Model
 ): WebSearchPluginConfig | undefined {
  switch (providerId) {
+    case 'azure-responses':
    case 'openai': {
      const searchContextSize = isOpenAIDeepResearchModel(model)
        ? 'medium'
@@ -1,35 +1,120 @@
+import 'emoji-picker-element'
+
 import TwemojiCountryFlagsWoff2 from '@renderer/assets/fonts/country-flag-fonts/TwemojiCountryFlags.woff2?url'
 import { useTheme } from '@renderer/context/ThemeProvider'
+import type { LanguageVarious } from '@renderer/types'
 import { polyfillCountryFlagEmojis } from 'country-flag-emoji-polyfill'
+// i18n translations from emoji-picker-element
+import de from 'emoji-picker-element/i18n/de'
+import en from 'emoji-picker-element/i18n/en'
+import es from 'emoji-picker-element/i18n/es'
+import fr from 'emoji-picker-element/i18n/fr'
+import ja from 'emoji-picker-element/i18n/ja'
+import pt_PT from 'emoji-picker-element/i18n/pt_PT'
+import ru_RU from 'emoji-picker-element/i18n/ru_RU'
+import zh_CN from 'emoji-picker-element/i18n/zh_CN'
+import type Picker from 'emoji-picker-element/picker'
+import type { EmojiClickEvent, NativeEmoji } from 'emoji-picker-element/shared'
+// Emoji data from emoji-picker-element-data (local, no CDN)
+// Using CLDR format for full multi-language search support (28 languages)
+import dataDE from 'emoji-picker-element-data/de/cldr/data.json?url'
+import dataEN from 'emoji-picker-element-data/en/cldr/data.json?url'
+import dataES from 'emoji-picker-element-data/es/cldr/data.json?url'
+import dataFR from 'emoji-picker-element-data/fr/cldr/data.json?url'
+import dataJA from 'emoji-picker-element-data/ja/cldr/data.json?url'
+import dataPT from 'emoji-picker-element-data/pt/cldr/data.json?url'
+import dataRU from 'emoji-picker-element-data/ru/cldr/data.json?url'
+import dataZH from 'emoji-picker-element-data/zh/cldr/data.json?url'
+import dataZH_HANT from 'emoji-picker-element-data/zh-hant/cldr/data.json?url'
 import type { FC } from 'react'
 import { useEffect, useRef } from 'react'
+import { useTranslation } from 'react-i18next'

 interface Props {
  onEmojiClick: (emoji: string) => void
 }

+// Mapping from app locale to emoji-picker-element i18n
+const i18nMap: Record<LanguageVarious, typeof en> = {
+  'en-US': en,
+  'zh-CN': zh_CN,
+  'zh-TW': zh_CN, // Closest available
+  'de-DE': de,
+  'el-GR': en, // No Greek available, fallback to English
+  'es-ES': es,
+  'fr-FR': fr,
+  'ja-JP': ja,
+  'pt-PT': pt_PT,
+  'ru-RU': ru_RU
+}
+
+// Mapping from app locale to emoji data URL
+// Using CLDR format provides native language search support for all locales
+const dataSourceMap: Record<LanguageVarious, string> = {
+  'en-US': dataEN,
+  'zh-CN': dataZH,
+  'zh-TW': dataZH_HANT,
+  'de-DE': dataDE,
+  'el-GR': dataEN, // No Greek CLDR available, fallback to English
+  'es-ES': dataES,
+  'fr-FR': dataFR,
+  'ja-JP': dataJA,
+  'pt-PT': dataPT,
+  'ru-RU': dataRU
+}
+
+// Mapping from app locale to emoji-picker-element locale string
+// Must match the data source locale for proper IndexedDB caching
+const localeMap: Record<LanguageVarious, string> = {
+  'en-US': 'en',
+  'zh-CN': 'zh',
+  'zh-TW': 'zh-hant',
+  'de-DE': 'de',
+  'el-GR': 'en',
+  'es-ES': 'es',
+  'fr-FR': 'fr',
+  'ja-JP': 'ja',
+  'pt-PT': 'pt',
+  'ru-RU': 'ru'
+}
+
 const EmojiPicker: FC<Props> = ({ onEmojiClick }) => {
  const { theme } = useTheme()
-  const ref = useRef<HTMLDivElement>(null)
+  const { i18n } = useTranslation()
+  const ref = useRef<Picker>(null)
+  const currentLocale = i18n.language as LanguageVarious

  useEffect(() => {
    polyfillCountryFlagEmojis('Twemoji Mozilla', TwemojiCountryFlagsWoff2)
  }, [])

+  // Configure picker with i18n and dataSource
  useEffect(() => {
-    const refValue = ref.current
+    const picker = ref.current
+    if (picker) {
+      picker.i18n = i18nMap[currentLocale] || en
+      picker.dataSource = dataSourceMap[currentLocale] || dataEN
+      picker.locale = localeMap[currentLocale] || 'en'
+    }
+  }, [currentLocale])

-    if (refValue) {
-      const handleEmojiClick = (event: any) => {
+  useEffect(() => {
+    const picker = ref.current
+
+    if (picker) {
+      const handleEmojiClick = (event: EmojiClickEvent) => {
        event.stopPropagation()
-        onEmojiClick(event.detail.unicode || event.detail.emoji.unicode)
+        const { detail } = event
+        // Use detail.unicode (processed with skin tone) or fallback to emoji's unicode for native emoji
+        const unicode = detail.unicode || ('unicode' in detail.emoji ? (detail.emoji as NativeEmoji).unicode : '')
+        onEmojiClick(unicode)
      }
      // 添加事件监听器
-      refValue.addEventListener('emoji-click', handleEmojiClick)
+      picker.addEventListener('emoji-click', handleEmojiClick)

      // 清理事件监听器
      return () => {
-        refValue.removeEventListener('emoji-click', handleEmojiClick)
+        picker.removeEventListener('emoji-click', handleEmojiClick)
      }
    }
    return
@@ -0,0 +1,157 @@
+import { loggerService } from '@logger'
+import type { UIActionResult } from '@mcp-ui/client'
+import { UIResourceRenderer } from '@mcp-ui/client'
+import type { EmbeddedResource } from '@modelcontextprotocol/sdk/types.js'
+import { isUIResource } from '@renderer/types'
+import type { FC } from 'react'
+import { useCallback, useState } from 'react'
+import { useTranslation } from 'react-i18next'
+import styled from 'styled-components'
+
+const logger = loggerService.withContext('MCPUIRenderer')
+
+interface Props {
+  resource: EmbeddedResource
+  serverId?: string
+  serverName?: string
+  onToolCall?: (toolName: string, params: any) => Promise<any>
+}
+
+const MCPUIRenderer: FC<Props> = ({ resource, onToolCall }) => {
+  const { t } = useTranslation()
+  const [error] = useState<string | null>(null)
+
+  const handleUIAction = useCallback(
+    async (result: UIActionResult): Promise<any> => {
+      logger.debug('UI Action received:', result)
+
+      try {
+        switch (result.type) {
+          case 'tool': {
+            // Handle tool call from UI
+            if (onToolCall) {
+              const { toolName, params } = result.payload
+              logger.info(`UI requesting tool call: ${toolName}`, { params })
+              const response = await onToolCall(toolName, params)
+
+              // Check if the response contains a UIResource
+              try {
+                if (response && response.content && Array.isArray(response.content)) {
+                  const firstContent = response.content[0]
+                  if (firstContent && firstContent.type === 'text' && firstContent.text) {
+                    const parsedText = JSON.parse(firstContent.text)
+                    if (isUIResource(parsedText)) {
+                      // Return the UIResource directly for rendering in the iframe
+                      logger.info('Tool response contains UIResource:', { uri: parsedText.resource.uri })
+                      return { status: 'success', data: parsedText }
+                    }
+                  }
+                }
+              } catch (parseError) {
+                // Not a UIResource, return the original response
+                logger.debug('Tool response is not a UIResource')
+              }
+
+              return { status: 'success', data: response }
+            } else {
+              logger.warn('Tool call requested but no handler provided')
+              return { status: 'error', message: 'Tool call handler not available' }
+            }
+          }
+
+          case 'intent': {
+            // Handle user intent
+            logger.info('UI intent:', result.payload)
+            window.toast.info(t('message.mcp.ui.intent_received'))
+            return { status: 'acknowledged' }
+          }
+
+          case 'notify': {
+            // Handle notification from UI
+            logger.info('UI notification:', result.payload)
+            window.toast.info(result.payload.message || t('message.mcp.ui.notification'))
+            return { status: 'acknowledged' }
+          }
+
+          case 'prompt': {
+            // Handle prompt request from UI
+            logger.info('UI prompt request:', result.payload)
+            // TODO: Integrate with prompt system
+            return { status: 'error', message: 'Prompt execution not yet implemented' }
+          }
+
+          case 'link': {
+            // Handle navigation request
+            const { url } = result.payload
+            logger.info('UI navigation request:', { url })
+            window.open(url, '_blank')
+            return { status: 'acknowledged' }
+          }
+
+          default:
+            logger.warn('Unknown UI action type:', { result })
+            return { status: 'error', message: 'Unknown action type' }
+        }
+      } catch (err) {
+        logger.error('Error handling UI action:', err as Error)
+        return {
+          status: 'error',
+          message: err instanceof Error ? err.message : 'Unknown error'
+        }
+      }
+    },
+    [onToolCall, t]
+  )
+
+  if (error) {
+    return (
+      <ErrorContainer>
+        <ErrorTitle>{t('message.mcp.ui.error')}</ErrorTitle>
+        <ErrorMessage>{error}</ErrorMessage>
+      </ErrorContainer>
+    )
+  }
+
+  return (
+    <UIContainer>
+      <UIResourceRenderer resource={resource} onUIAction={handleUIAction} />
+    </UIContainer>
+  )
+}
+
+const UIContainer = styled.div`
+  width: 100%;
+  min-height: 400px;
+  border-radius: 8px;
+  overflow: hidden;
+  background: var(--color-background);
+  border: 1px solid var(--color-border);
+
+  iframe {
+    width: 100%;
+    border: none;
+    min-height: 400px;
+    height: 600px;
+  }
+`
+
+const ErrorContainer = styled.div`
+  padding: 16px;
+  border-radius: 8px;
+  background: var(--color-error-bg, #fee);
+  border: 1px solid var(--color-error-border, #fcc);
+  color: var(--color-error-text, #c33);
+`
+
+const ErrorTitle = styled.div`
+  font-weight: 600;
+  margin-bottom: 8px;
+  font-size: 14px;
+`
+
+const ErrorMessage = styled.div`
+  font-size: 13px;
+  opacity: 0.9;
+`
+
+export default MCPUIRenderer
@@ -0,0 +1 @@
+export { default as MCPUIRenderer } from './MCPUIRenderer'
@@ -0,0 +1,141 @@
+import { importChatGPTConversations } from '@renderer/services/import'
+import { Alert, Modal, Progress, Space, Spin } from 'antd'
+import { useState } from 'react'
+import { useTranslation } from 'react-i18next'
+
+import { TopView } from '../TopView'
+
+interface PopupResult {
+  success?: boolean
+}
+
+interface Props {
+  resolve: (data: PopupResult) => void
+}
+
+const PopupContainer: React.FC<Props> = ({ resolve }) => {
+  const [open, setOpen] = useState(true)
+  const [selecting, setSelecting] = useState(false)
+  const [importing, setImporting] = useState(false)
+  const { t } = useTranslation()
+
+  const onOk = async () => {
+    setSelecting(true)
+    try {
+      // Select ChatGPT JSON file
+      const file = await window.api.file.open({
+        filters: [{ name: 'ChatGPT Conversations', extensions: ['json'] }]
+      })
+
+      setSelecting(false)
+
+      if (!file) {
+        return
+      }
+
+      setImporting(true)
+
+      // Parse file content
+      const fileContent = typeof file.content === 'string' ? file.content : new TextDecoder().decode(file.content)
+
+      // Import conversations
+      const result = await importChatGPTConversations(fileContent)
+
+      if (result.success) {
+        window.toast.success(
+          t('import.chatgpt.success', {
+            topics: result.topicsCount,
+            messages: result.messagesCount
+          })
+        )
+        setOpen(false)
+      } else {
+        window.toast.error(result.error || t('import.chatgpt.error.unknown'))
+      }
+    } catch (error) {
+      window.toast.error(t('import.chatgpt.error.unknown'))
+      setOpen(false)
+    } finally {
+      setSelecting(false)
+      setImporting(false)
+    }
+  }
+
+  const onCancel = () => {
+    setOpen(false)
+  }
+
+  const onClose = () => {
+    resolve({})
+  }
+
+  ImportPopup.hide = onCancel
+
+  return (
+    <Modal
+      title={t('import.chatgpt.title')}
+      open={open}
+      onOk={onOk}
+      onCancel={onCancel}
+      afterClose={onClose}
+      okText={t('import.chatgpt.button')}
+      okButtonProps={{ disabled: selecting || importing, loading: selecting }}
+      cancelButtonProps={{ disabled: selecting || importing }}
+      maskClosable={false}
+      transitionName="animation-move-down"
+      centered>
+      {!selecting && !importing && (
+        <Space direction="vertical" style={{ width: '100%' }}>
+          <div>{t('import.chatgpt.description')}</div>
+          <Alert
+            message={t('import.chatgpt.help.title')}
+            description={
+              <div>
+                <p>{t('import.chatgpt.help.step1')}</p>
+                <p>{t('import.chatgpt.help.step2')}</p>
+                <p>{t('import.chatgpt.help.step3')}</p>
+              </div>
+            }
+            type="info"
+            showIcon
+            style={{ marginTop: 12 }}
+          />
+        </Space>
+      )}
+      {selecting && (
+        <div style={{ textAlign: 'center', padding: '40px 0' }}>
+          <Spin size="large" />
+          <div style={{ marginTop: 16 }}>{t('import.chatgpt.selecting')}</div>
+        </div>
+      )}
+      {importing && (
+        <div style={{ textAlign: 'center', padding: '20px 0' }}>
+          <Progress percent={100} status="active" strokeColor="var(--color-primary)" showInfo={false} />
+          <div style={{ marginTop: 16 }}>{t('import.chatgpt.importing')}</div>
+        </div>
+      )}
+    </Modal>
+  )
+}
+
+const TopViewKey = 'ImportPopup'
+
+export default class ImportPopup {
+  static topviewId = 0
+  static hide() {
+    TopView.hide(TopViewKey)
+  }
+  static show() {
+    return new Promise<PopupResult>((resolve) => {
+      TopView.show(
+        <PopupContainer
+          resolve={(v) => {
+            resolve(v)
+            TopView.hide(TopViewKey)
+          }}
+        />,
+        TopViewKey
+      )
+    })
+  }
+}
@@ -1,5 +1,4 @@
 import { loggerService } from '@logger'
-import ClaudeIcon from '@renderer/assets/images/models/claude.png'
 import { ErrorBoundary } from '@renderer/components/ErrorBoundary'
 import { TopView } from '@renderer/components/TopView'
 import { permissionModeCards } from '@renderer/config/agent'
@@ -9,7 +8,6 @@ import SelectAgentBaseModelButton from '@renderer/pages/home/components/SelectAg
 import type {
  AddAgentForm,
  AgentEntity,
-  AgentType,
  ApiModel,
  BaseAgentForm,
  PermissionMode,
@@ -17,30 +15,22 @@ import type {
  UpdateAgentForm
 } from '@renderer/types'
 import { AgentConfigurationSchema, isAgentType } from '@renderer/types'
-import { Avatar, Button, Input, Modal, Select } from 'antd'
+import { Alert, Button, Input, Modal, Select } from 'antd'
 import { AlertTriangleIcon } from 'lucide-react'
 import type { ChangeEvent, FormEvent } from 'react'
 import { useCallback, useEffect, useMemo, useRef, useState } from 'react'
 import { useTranslation } from 'react-i18next'
 import styled from 'styled-components'

-import type { BaseOption } from './shared'
-
 const { TextArea } = Input

 const logger = loggerService.withContext('AddAgentPopup')

-interface AgentTypeOption extends BaseOption {
-  type: 'type'
-  key: AgentEntity['type']
-  name: AgentEntity['name']
-}
-
 type AgentWithTools = AgentEntity & { tools?: Tool[] }

 const buildAgentForm = (existing?: AgentWithTools): BaseAgentForm => ({
  type: existing?.type ?? 'claude-code',
-  name: existing?.name ?? 'Claude Code',
+  name: existing?.name ?? 'Agent',
  description: existing?.description,
  instructions: existing?.instructions,
  model: existing?.model ?? '',
@@ -68,6 +58,7 @@ const PopupContainer: React.FC<Props> = ({ agent, afterSubmit, resolve }) => {
  const isEditing = (agent?: AgentWithTools) => agent !== undefined

  const [form, setForm] = useState<BaseAgentForm>(() => buildAgentForm(agent))
+  const [hasGitBash, setHasGitBash] = useState<boolean>(true)

  useEffect(() => {
    if (open) {
@@ -75,6 +66,30 @@ const PopupContainer: React.FC<Props> = ({ agent, afterSubmit, resolve }) => {
    }
  }, [agent, open])

+  const checkGitBash = useCallback(
+    async (showToast = false) => {
+      try {
+        const gitBashInstalled = await window.api.system.checkGitBash()
+        setHasGitBash(gitBashInstalled)
+        if (showToast) {
+          if (gitBashInstalled) {
+            window.toast.success(t('agent.gitBash.success', 'Git Bash detected successfully!'))
+          } else {
+            window.toast.error(t('agent.gitBash.notFound', 'Git Bash not found. Please install it first.'))
+          }
+        }
+      } catch (error) {
+        logger.error('Failed to check Git Bash:', error as Error)
+        setHasGitBash(true) // Default to true on error to avoid false warnings
+      }
+    },
+    [t]
+  )
+
+  useEffect(() => {
+    checkGitBash()
+  }, [checkGitBash])
+
  const selectedPermissionMode = form.configuration?.permission_mode ?? 'default'

  const onPermissionModeChange = useCallback((value: PermissionMode) => {
@@ -100,54 +115,6 @@ const PopupContainer: React.FC<Props> = ({ agent, afterSubmit, resolve }) => {
    })
  }, [])

-  // add supported agents type here.
-  const agentConfig = useMemo(
-    () =>
-      [
-        {
-          type: 'type',
-          key: 'claude-code',
-          label: 'Claude Code',
-          name: 'Claude Code',
-          avatar: ClaudeIcon
-        }
-      ] as const satisfies AgentTypeOption[],
-    []
-  )
-
-  const agentOptions = useMemo(
-    () =>
-      agentConfig.map((option) => ({
-        value: option.key,
-        label: (
-          <OptionWrapper>
-            <Avatar src={option.avatar} size={24} />
-            <span>{option.label}</span>
-          </OptionWrapper>
-        )
-      })),
-    [agentConfig]
-  )
-
-  const onAgentTypeChange = useCallback(
-    (value: AgentType) => {
-      const prevConfig = agentConfig.find((config) => config.key === form.type)
-      let newName: string | undefined = form.name
-      if (prevConfig && prevConfig.name === form.name) {
-        const newConfig = agentConfig.find((config) => config.key === value)
-        if (newConfig) {
-          newName = newConfig.name
-        }
-      }
-      setForm((prev) => ({
-        ...prev,
-        type: value,
-        name: newName
-      }))
-    },
-    [agentConfig, form.name, form.type]
-  )
-
  const onNameChange = useCallback((e: ChangeEvent<HTMLInputElement>) => {
    setForm((prev) => ({
      ...prev,
@@ -155,12 +122,12 @@ const PopupContainer: React.FC<Props> = ({ agent, afterSubmit, resolve }) => {
    }))
  }, [])

-  const onDescChange = useCallback((e: ChangeEvent<HTMLTextAreaElement>) => {
-    setForm((prev) => ({
-      ...prev,
-      description: e.target.value
-    }))
-  }, [])
+  // const onDescChange = useCallback((e: ChangeEvent<HTMLTextAreaElement>) => {
+  //   setForm((prev) => ({
+  //     ...prev,
+  //     description: e.target.value
+  //   }))
+  // }, [])

  const onInstChange = useCallback((e: ChangeEvent<HTMLTextAreaElement>) => {
    setForm((prev) => ({
@@ -333,17 +300,37 @@ const PopupContainer: React.FC<Props> = ({ agent, afterSubmit, resolve }) => {
        footer={null}>
        <StyledForm onSubmit={onSubmit}>
          <FormContent>
+            {!hasGitBash && (
+              <Alert
+                message={t('agent.gitBash.error.title', 'Git Bash Required')}
+                description={
+                  <div>
+                    <div style={{ marginBottom: 8 }}>
+                      {t(
+                        'agent.gitBash.error.description',
+                        'Git Bash is required to run agents on Windows. The agent cannot function without it. Please install Git for Windows from'
+                      )}{' '}
+                      <a
+                        href="https://git-scm.com/download/win"
+                        onClick={(e) => {
+                          e.preventDefault()
+                          window.api.openWebsite('https://git-scm.com/download/win')
+                        }}
+                        style={{ textDecoration: 'underline' }}>
+                        git-scm.com
+                      </a>
+                    </div>
+                    <Button size="small" onClick={() => checkGitBash(true)}>
+                      {t('agent.gitBash.error.recheck', 'Recheck Git Bash Installation')}
+                    </Button>
+                  </div>
+                }
+                type="error"
+                showIcon
+                style={{ marginBottom: 16 }}
+              />
+            )}
            <FormRow>
-              <FormItem style={{ flex: 1 }}>
-                <Label>{t('agent.type.label')}</Label>
-                <Select
-                  value={form.type}
-                  onChange={onAgentTypeChange}
-                  options={agentOptions}
-                  disabled={isEditing(agent)}
-                  style={{ width: '100%' }}
-                />
-              </FormItem>
              <FormItem style={{ flex: 1 }}>
                <Label>
                  {t('common.name')} <RequiredMark>*</RequiredMark>
@@ -363,7 +350,7 @@ const PopupContainer: React.FC<Props> = ({ agent, afterSubmit, resolve }) => {
                avatarSize={24}
                iconSize={16}
                buttonStyle={{
-                  padding: '8px 12px',
+                  padding: '3px 8px',
                  width: '100%',
                  border: '1px solid var(--color-border)',
                  borderRadius: 6,
@@ -382,7 +369,6 @@ const PopupContainer: React.FC<Props> = ({ agent, afterSubmit, resolve }) => {
                onChange={onPermissionModeChange}
                style={{ width: '100%' }}
                placeholder={t('agent.settings.tooling.permissionMode.placeholder', 'Select permission mode')}
-                dropdownStyle={{ minWidth: '500px' }}
                optionLabelProp="label">
                {permissionModeCards.map((item) => (
                  <Select.Option key={item.mode} value={item.mode} label={t(item.titleKey, item.titleFallback)}>
@@ -438,15 +424,15 @@ const PopupContainer: React.FC<Props> = ({ agent, afterSubmit, resolve }) => {
              <TextArea rows={3} value={form.instructions ?? ''} onChange={onInstChange} />
            </FormItem>

-            <FormItem>
+            {/* <FormItem>
              <Label>{t('common.description')}</Label>
-              <TextArea rows={2} value={form.description ?? ''} onChange={onDescChange} />
-            </FormItem>
+              <TextArea rows={1} value={form.description ?? ''} onChange={onDescChange} />
+            </FormItem> */}
          </FormContent>

          <FormFooter>
            <Button onClick={onCancel}>{t('common.close')}</Button>
-            <Button type="primary" htmlType="submit" loading={loadingRef.current}>
+            <Button type="primary" htmlType="submit" loading={loadingRef.current} disabled={!hasGitBash}>
              {isEditing(agent) ? t('common.confirm') : t('common.add')}
            </Button>
          </FormFooter>
@@ -575,14 +561,7 @@ const FormFooter = styled.div`
  display: flex;
  justify-content: flex-end;
  gap: 8px;
-  padding-top: 16px;
-  border-top: 1px solid var(--color-border);
-`
-
-const OptionWrapper = styled.div`
-  display: flex;
-  align-items: center;
-  gap: 8px;
+  padding: 10px;
 `

 const PermissionOptionWrapper = styled.div`
@@ -6,7 +6,7 @@ import { useEffect, useMemo, useRef, useState } from 'react'
 import { useTranslation } from 'react-i18next'
 import styled, { css } from 'styled-components'

-interface SelectorOption<V = string | number> {
+interface SelectorOption<V = string | number | undefined | null> {
  label: string | ReactNode
  value: V
  type?: 'group'
@@ -14,7 +14,7 @@ interface SelectorOption<V = string | number> {
  disabled?: boolean
 }

-interface BaseSelectorProps<V = string | number> {
+interface BaseSelectorProps<V = string | number | undefined | null> {
  options: SelectorOption<V>[]
  placeholder?: string
  placement?: 'topLeft' | 'topCenter' | 'topRight' | 'bottomLeft' | 'bottomCenter' | 'bottomRight' | 'top' | 'bottom'
@@ -39,7 +39,7 @@ interface MultipleSelectorProps<V> extends BaseSelectorProps<V> {

 export type SelectorProps<V> = SingleSelectorProps<V> | MultipleSelectorProps<V>

-const Selector = <V extends string | number>({
+const Selector = <V extends string | number | undefined | null>({
  options,
  value,
  onChange = () => {},
@@ -1,233 +0,0 @@
-import { describe, expect, it, vi } from 'vitest'
-
-import { isDoubaoSeedAfter251015, isDoubaoThinkingAutoModel, isLingReasoningModel } from '../models/reasoning'
-
-vi.mock('@renderer/store', () => ({
-  default: {
-    getState: () => ({
-      llm: {
-        settings: {}
-      }
-    })
-  }
-}))
-
-// FIXME: Idk why it's imported. Maybe circular dependency somewhere
-vi.mock('@renderer/services/AssistantService.ts', () => ({
-  getDefaultAssistant: () => {
-    return {
-      id: 'default',
-      name: 'default',
-      emoji: '😀',
-      prompt: '',
-      topics: [],
-      messages: [],
-      type: 'assistant',
-      regularPhrases: [],
-      settings: {}
-    }
-  }
-}))
-
-describe('Doubao Models', () => {
-  describe('isDoubaoThinkingAutoModel', () => {
-    it('should return false for invalid models', () => {
-      expect(
-        isDoubaoThinkingAutoModel({
-          id: 'doubao-seed-1-6-251015',
-          name: 'doubao-seed-1-6-251015',
-          provider: '',
-          group: ''
-        })
-      ).toBe(false)
-      expect(
-        isDoubaoThinkingAutoModel({
-          id: 'doubao-seed-1-6-lite-251015',
-          name: 'doubao-seed-1-6-lite-251015',
-          provider: '',
-          group: ''
-        })
-      ).toBe(false)
-      expect(
-        isDoubaoThinkingAutoModel({
-          id: 'doubao-seed-1-6-thinking-250715',
-          name: 'doubao-seed-1-6-thinking-250715',
-          provider: '',
-          group: ''
-        })
-      ).toBe(false)
-      expect(
-        isDoubaoThinkingAutoModel({
-          id: 'doubao-seed-1-6-flash',
-          name: 'doubao-seed-1-6-flash',
-          provider: '',
-          group: ''
-        })
-      ).toBe(false)
-      expect(
-        isDoubaoThinkingAutoModel({
-          id: 'doubao-seed-1-6-thinking',
-          name: 'doubao-seed-1-6-thinking',
-          provider: '',
-          group: ''
-        })
-      ).toBe(false)
-    })
-
-    it('should return true for valid models', () => {
-      expect(
-        isDoubaoThinkingAutoModel({
-          id: 'doubao-seed-1-6-250615',
-          name: 'doubao-seed-1-6-250615',
-          provider: '',
-          group: ''
-        })
-      ).toBe(true)
-      expect(
-        isDoubaoThinkingAutoModel({
-          id: 'Doubao-Seed-1.6',
-          name: 'Doubao-Seed-1.6',
-          provider: '',
-          group: ''
-        })
-      ).toBe(true)
-      expect(
-        isDoubaoThinkingAutoModel({
-          id: 'doubao-1-5-thinking-pro-m',
-          name: 'doubao-1-5-thinking-pro-m',
-          provider: '',
-          group: ''
-        })
-      ).toBe(true)
-      expect(
-        isDoubaoThinkingAutoModel({
-          id: 'doubao-seed-1.6-lite',
-          name: 'doubao-seed-1.6-lite',
-          provider: '',
-          group: ''
-        })
-      ).toBe(true)
-      expect(
-        isDoubaoThinkingAutoModel({
-          id: 'doubao-1-5-thinking-pro-m-12345',
-          name: 'doubao-1-5-thinking-pro-m-12345',
-          provider: '',
-          group: ''
-        })
-      ).toBe(true)
-    })
-  })
-
-  describe('isDoubaoSeedAfter251015', () => {
-    it('should return true for models matching the pattern', () => {
-      expect(
-        isDoubaoSeedAfter251015({
-          id: 'doubao-seed-1-6-251015',
-          name: '',
-          provider: '',
-          group: ''
-        })
-      ).toBe(true)
-      expect(
-        isDoubaoSeedAfter251015({
-          id: 'doubao-seed-1-6-lite-251015',
-          name: '',
-          provider: '',
-          group: ''
-        })
-      ).toBe(true)
-    })
-
-    it('should return false for models not matching the pattern', () => {
-      expect(
-        isDoubaoSeedAfter251015({
-          id: 'doubao-seed-1-6-250615',
-          name: '',
-          provider: '',
-          group: ''
-        })
-      ).toBe(false)
-      expect(
-        isDoubaoSeedAfter251015({
-          id: 'Doubao-Seed-1.6',
-          name: '',
-          provider: '',
-          group: ''
-        })
-      ).toBe(false)
-      expect(
-        isDoubaoSeedAfter251015({
-          id: 'doubao-1-5-thinking-pro-m',
-          name: '',
-          provider: '',
-          group: ''
-        })
-      ).toBe(false)
-      expect(
-        isDoubaoSeedAfter251015({
-          id: 'doubao-seed-1-6-lite-251016',
-          name: '',
-          provider: '',
-          group: ''
-        })
-      ).toBe(false)
-    })
-  })
-})
-describe('Ling Models', () => {
-  describe('isLingReasoningModel', () => {
-    it('should return false for ling variants', () => {
-      expect(
-        isLingReasoningModel({
-          id: 'ling-1t',
-          name: '',
-          provider: '',
-          group: ''
-        })
-      ).toBe(false)
-      expect(
-        isLingReasoningModel({
-          id: 'ling-flash-2.0',
-          name: '',
-          provider: '',
-          group: ''
-        })
-      ).toBe(false)
-      expect(
-        isLingReasoningModel({
-          id: 'ling-mini-2.0',
-          name: '',
-          provider: '',
-          group: ''
-        })
-      ).toBe(false)
-    })
-
-    it('should return true for ring variants', () => {
-      expect(
-        isLingReasoningModel({
-          id: 'ring-1t',
-          name: '',
-          provider: '',
-          group: ''
-        })
-      ).toBe(true)
-      expect(
-        isLingReasoningModel({
-          id: 'ring-flash-2.0',
-          name: '',
-          provider: '',
-          group: ''
-        })
-      ).toBe(true)
-      expect(
-        isLingReasoningModel({
-          id: 'ring-mini-2.0',
-          name: '',
-          provider: '',
-          group: ''
-        })
-      ).toBe(true)
-    })
-  })
-})
@@ -0,0 +1,101 @@
+import type { Model } from '@renderer/types'
+import { describe, expect, it, vi } from 'vitest'
+
+vi.mock('@renderer/hooks/useStore', () => ({
+  getStoreProviders: vi.fn(() => [])
+}))
+
+vi.mock('@renderer/store', () => ({
+  __esModule: true,
+  default: {
+    getState: () => ({
+      llm: { providers: [] },
+      settings: {}
+    })
+  },
+  useAppDispatch: vi.fn(),
+  useAppSelector: vi.fn()
+}))
+
+vi.mock('@renderer/store/settings', () => {
+  const noop = vi.fn()
+  return new Proxy(
+    {},
+    {
+      get: (_target, prop) => {
+        if (prop === 'initialState') {
+          return {}
+        }
+        return noop
+      }
+    }
+  )
+})
+
+vi.mock('@renderer/hooks/useSettings', () => ({
+  useSettings: vi.fn(() => ({})),
+  useNavbarPosition: vi.fn(() => ({ navbarPosition: 'left' })),
+  useMessageStyle: vi.fn(() => ({ isBubbleStyle: false })),
+  getStoreSetting: vi.fn()
+}))
+
+import { isEmbeddingModel, isRerankModel } from '../embedding'
+
+const createModel = (overrides: Partial<Model> = {}): Model => ({
+  id: 'test-model',
+  name: 'Test Model',
+  provider: 'openai',
+  group: 'Test',
+  ...overrides
+})
+
+describe('isEmbeddingModel', () => {
+  it('returns true for ids that match the embedding regex', () => {
+    expect(isEmbeddingModel(createModel({ id: 'Text-Embedding-3-Small' }))).toBe(true)
+  })
+
+  it('returns false for rerank models even if they match embedding patterns', () => {
+    const model = createModel({ id: 'rerank-qa', name: 'rerank-qa' })
+    expect(isRerankModel(model)).toBe(true)
+    expect(isEmbeddingModel(model)).toBe(false)
+  })
+
+  it('honors user overrides for embedding capability', () => {
+    const model = createModel({
+      id: 'text-embedding-3-small',
+      capabilities: [{ type: 'embedding', isUserSelected: false }]
+    })
+    expect(isEmbeddingModel(model)).toBe(false)
+  })
+
+  it('uses the model name when provider is doubao', () => {
+    const model = createModel({
+      id: 'custom-id',
+      name: 'BGE-Large-zh-v1.5',
+      provider: 'doubao'
+    })
+    expect(isEmbeddingModel(model)).toBe(true)
+  })
+
+  it('returns false for anthropic provider models', () => {
+    const model = createModel({
+      id: 'text-embedding-ada-002',
+      provider: 'anthropic'
+    })
+    expect(isEmbeddingModel(model)).toBe(false)
+  })
+})
+
+describe('isRerankModel', () => {
+  it('identifies ids that match rerank regex', () => {
+    expect(isRerankModel(createModel({ id: 'jina-rerank-v2-base' }))).toBe(true)
+  })
+
+  it('honors user overrides for rerank capability', () => {
+    const model = createModel({
+      id: 'jina-rerank-v2-base',
+      capabilities: [{ type: 'rerank', isUserSelected: false }]
+    })
+    expect(isRerankModel(model)).toBe(false)
+  })
+})
@@ -1,33 +1,55 @@
 import {
  isImageEnhancementModel,
-  isPureGenerateImageModel,
  isQwenReasoningModel,
  isSupportedThinkingTokenQwenModel,
-  isVisionModel,
-  isWebSearchModel
+  isVisionModel
 } from '@renderer/config/models'
 import type { Model } from '@renderer/types'
 import { beforeEach, describe, expect, test, vi } from 'vitest'

+vi.mock('@renderer/store/llm', () => ({
+  initialState: {}
+}))
+
+vi.mock('@renderer/store', () => ({
+  default: {
+    getState: () => ({
+      llm: {
+        settings: {}
+      }
+    })
+  }
+}))
+
+const getProviderByModelMock = vi.fn()
+const isEmbeddingModelMock = vi.fn()
+const isRerankModelMock = vi.fn()
+
+vi.mock('@renderer/services/AssistantService', () => ({
+  getProviderByModel: (...args: any[]) => getProviderByModelMock(...args),
+  getAssistantSettings: vi.fn(),
+  getDefaultAssistant: vi.fn().mockReturnValue({
+    id: 'default',
+    name: 'Default Assistant',
+    prompt: '',
+    settings: {}
+  })
+}))
+
+vi.mock('@renderer/config/models/embedding', () => ({
+  isEmbeddingModel: (...args: any[]) => isEmbeddingModelMock(...args),
+  isRerankModel: (...args: any[]) => isRerankModelMock(...args)
+}))
+
+beforeEach(() => {
+  vi.clearAllMocks()
+  getProviderByModelMock.mockReturnValue({ type: 'openai-response' } as any)
+  isEmbeddingModelMock.mockReturnValue(false)
+  isRerankModelMock.mockReturnValue(false)
+})
+
 // Suggested test cases
 describe('Qwen Model Detection', () => {
-  beforeEach(() => {
-    vi.mock('@renderer/store/llm', () => ({
-      initialState: {}
-    }))
-    vi.mock('@renderer/services/AssistantService', () => ({
-      getProviderByModel: vi.fn().mockReturnValue({ id: 'cherryai' })
-    }))
-    vi.mock('@renderer/store', () => ({
-      default: {
-        getState: () => ({
-          llm: {
-            settings: {}
-          }
-        })
-      }
-    }))
-  })
  test('isQwenReasoningModel', () => {
    expect(isQwenReasoningModel({ id: 'qwen3-thinking' } as Model)).toBe(true)
    expect(isQwenReasoningModel({ id: 'qwen3-instruct' } as Model)).toBe(false)
@@ -56,14 +78,6 @@ describe('Qwen Model Detection', () => {
 })

 describe('Vision Model Detection', () => {
-  beforeEach(() => {
-    vi.mock('@renderer/store/llm', () => ({
-      initialState: {}
-    }))
-    vi.mock('@renderer/services/AssistantService', () => ({
-      getProviderByModel: vi.fn().mockReturnValue({ id: 'cherryai' })
-    }))
-  })
  test('isVisionModel', () => {
    expect(isVisionModel({ id: 'qwen-vl-max' } as Model)).toBe(true)
    expect(isVisionModel({ id: 'qwen-omni-turbo' } as Model)).toBe(true)
@@ -75,25 +89,4 @@ describe('Vision Model Detection', () => {
    expect(isImageEnhancementModel({ id: 'qwen-image-edit' } as Model)).toBe(true)
    expect(isImageEnhancementModel({ id: 'grok-2-image-latest' } as Model)).toBe(true)
  })
-  test('isPureGenerateImageModel', () => {
-    expect(isPureGenerateImageModel({ id: 'gpt-image-1' } as Model)).toBe(true)
-    expect(isPureGenerateImageModel({ id: 'gemini-2.5-flash-image-preview' } as Model)).toBe(true)
-    expect(isPureGenerateImageModel({ id: 'gemini-2.0-flash-preview-image-generation' } as Model)).toBe(true)
-    expect(isPureGenerateImageModel({ id: 'grok-2-image-latest' } as Model)).toBe(true)
-    expect(isPureGenerateImageModel({ id: 'gpt-4o' } as Model)).toBe(false)
-  })
-})
-
-describe('Web Search Model Detection', () => {
-  beforeEach(() => {
-    vi.mock('@renderer/store/llm', () => ({
-      initialState: {}
-    }))
-    vi.mock('@renderer/services/AssistantService', () => ({
-      getProviderByModel: vi.fn().mockReturnValue({ id: 'cherryai' })
-    }))
-  })
-  test('isWebSearchModel', () => {
-    expect(isWebSearchModel({ id: 'grok-2-image-latest' } as Model)).toBe(false)
-  })
 })
@@ -0,0 +1,137 @@
+import type { Model } from '@renderer/types'
+import { beforeEach, describe, expect, it, vi } from 'vitest'
+
+import { isEmbeddingModel, isRerankModel } from '../embedding'
+import { isDeepSeekHybridInferenceModel } from '../reasoning'
+import { isFunctionCallingModel } from '../tooluse'
+import { isPureGenerateImageModel, isTextToImageModel } from '../vision'
+
+vi.mock('@renderer/hooks/useStore', () => ({
+  getStoreProviders: vi.fn(() => [])
+}))
+
+vi.mock('@renderer/store', () => ({
+  __esModule: true,
+  default: {
+    getState: () => ({
+      llm: { providers: [] },
+      settings: {}
+    })
+  },
+  useAppDispatch: vi.fn(),
+  useAppSelector: vi.fn()
+}))
+
+vi.mock('@renderer/store/settings', () => {
+  const noop = vi.fn()
+  return new Proxy(
+    {},
+    {
+      get: (_target, prop) => {
+        if (prop === 'initialState') {
+          return {}
+        }
+        return noop
+      }
+    }
+  )
+})
+
+vi.mock('@renderer/hooks/useSettings', () => ({
+  useSettings: vi.fn(() => ({})),
+  useNavbarPosition: vi.fn(() => ({ navbarPosition: 'left' })),
+  useMessageStyle: vi.fn(() => ({ isBubbleStyle: false })),
+  getStoreSetting: vi.fn()
+}))
+
+vi.mock('../embedding', () => ({
+  isEmbeddingModel: vi.fn(),
+  isRerankModel: vi.fn()
+}))
+
+vi.mock('../vision', () => ({
+  isPureGenerateImageModel: vi.fn(),
+  isTextToImageModel: vi.fn()
+}))
+
+vi.mock('../reasoning', () => ({
+  isDeepSeekHybridInferenceModel: vi.fn()
+}))
+
+const createModel = (overrides: Partial<Model> = {}): Model => ({
+  id: 'gpt-4o',
+  name: 'gpt-4o',
+  provider: 'openai',
+  group: 'OpenAI',
+  ...overrides
+})
+
+const embeddingMock = vi.mocked(isEmbeddingModel)
+const rerankMock = vi.mocked(isRerankModel)
+const pureImageMock = vi.mocked(isPureGenerateImageModel)
+const textToImageMock = vi.mocked(isTextToImageModel)
+const deepSeekHybridMock = vi.mocked(isDeepSeekHybridInferenceModel)
+
+describe('isFunctionCallingModel', () => {
+  beforeEach(() => {
+    vi.clearAllMocks()
+    embeddingMock.mockReturnValue(false)
+    rerankMock.mockReturnValue(false)
+    pureImageMock.mockReturnValue(false)
+    textToImageMock.mockReturnValue(false)
+    deepSeekHybridMock.mockReturnValue(false)
+  })
+
+  it('returns false when the model is undefined', () => {
+    expect(isFunctionCallingModel(undefined as unknown as Model)).toBe(false)
+  })
+
+  it('returns false when model is classified as embedding/rerank/image', () => {
+    embeddingMock.mockReturnValueOnce(true)
+    expect(isFunctionCallingModel(createModel())).toBe(false)
+  })
+
+  it('respect manual user overrides', () => {
+    const model = createModel({
+      capabilities: [{ type: 'function_calling', isUserSelected: false }]
+    })
+    expect(isFunctionCallingModel(model)).toBe(false)
+    const enabled = createModel({
+      capabilities: [{ type: 'function_calling', isUserSelected: true }]
+    })
+    expect(isFunctionCallingModel(enabled)).toBe(true)
+  })
+
+  it('matches doubao models by name when regex applies', () => {
+    const doubao = createModel({
+      id: 'custom-model',
+      name: 'Doubao-Seed-1.6-251015',
+      provider: 'doubao'
+    })
+    expect(isFunctionCallingModel(doubao)).toBe(true)
+  })
+
+  it('returns true for regex matches on standard providers', () => {
+    expect(isFunctionCallingModel(createModel({ id: 'gpt-5' }))).toBe(true)
+  })
+
+  it('excludes explicitly blocked ids', () => {
+    expect(isFunctionCallingModel(createModel({ id: 'gemini-1.5-flash' }))).toBe(false)
+  })
+
+  it('forces support for trusted providers', () => {
+    for (const provider of ['deepseek', 'anthropic', 'kimi', 'moonshot']) {
+      expect(isFunctionCallingModel(createModel({ provider }))).toBe(true)
+    }
+  })
+
+  it('returns true when identified as deepseek hybrid inference model', () => {
+    deepSeekHybridMock.mockReturnValueOnce(true)
+    expect(isFunctionCallingModel(createModel({ id: 'deepseek-v3-1', provider: 'custom' }))).toBe(true)
+  })
+
+  it('returns false for deepseek hybrid models behind restricted system providers', () => {
+    deepSeekHybridMock.mockReturnValueOnce(true)
+    expect(isFunctionCallingModel(createModel({ id: 'deepseek-v3-1', provider: 'dashscope' }))).toBe(false)
+  })
+})
@@ -0,0 +1,280 @@
+import { isEmbeddingModel, isRerankModel } from '@renderer/config/models/embedding'
+import type { Model } from '@renderer/types'
+import { beforeEach, describe, expect, it, vi } from 'vitest'
+
+import {
+  isGPT5ProModel,
+  isGPT5SeriesModel,
+  isGPT5SeriesReasoningModel,
+  isGPT51SeriesModel,
+  isOpenAIChatCompletionOnlyModel,
+  isOpenAILLMModel,
+  isOpenAIModel,
+  isOpenAIOpenWeightModel,
+  isOpenAIReasoningModel,
+  isSupportVerbosityModel
+} from '../openai'
+import { isQwenMTModel } from '../qwen'
+import {
+  agentModelFilter,
+  getModelSupportedVerbosity,
+  groupQwenModels,
+  isAnthropicModel,
+  isGeminiModel,
+  isGemmaModel,
+  isGenerateImageModels,
+  isMaxTemperatureOneModel,
+  isNotSupportedTextDelta,
+  isNotSupportSystemMessageModel,
+  isNotSupportTemperatureAndTopP,
+  isSupportedFlexServiceTier,
+  isSupportedModel,
+  isSupportFlexServiceTierModel,
+  isVisionModels,
+  isZhipuModel
+} from '../utils'
+import { isGenerateImageModel, isTextToImageModel, isVisionModel } from '../vision'
+import { isOpenAIWebSearchChatCompletionOnlyModel } from '../websearch'
+
+vi.mock('@renderer/hooks/useStore', () => ({
+  getStoreProviders: vi.fn(() => [])
+}))
+
+vi.mock('@renderer/store', () => ({
+  __esModule: true,
+  default: {
+    getState: () => ({
+      llm: { providers: [] },
+      settings: {}
+    })
+  },
+  useAppDispatch: vi.fn(),
+  useAppSelector: vi.fn()
+}))
+
+vi.mock('@renderer/store/settings', () => {
+  const noop = vi.fn()
+  return new Proxy(
+    {},
+    {
+      get: (_target, prop) => {
+        if (prop === 'initialState') {
+          return {}
+        }
+        return noop
+      }
+    }
+  )
+})
+
+vi.mock('@renderer/hooks/useSettings', () => ({
+  useSettings: vi.fn(() => ({})),
+  useNavbarPosition: vi.fn(() => ({ navbarPosition: 'left' })),
+  useMessageStyle: vi.fn(() => ({ isBubbleStyle: false })),
+  getStoreSetting: vi.fn()
+}))
+
+vi.mock('@renderer/config/models/embedding', () => ({
+  isEmbeddingModel: vi.fn(),
+  isRerankModel: vi.fn()
+}))
+
+vi.mock('../vision', () => ({
+  isGenerateImageModel: vi.fn(),
+  isTextToImageModel: vi.fn(),
+  isVisionModel: vi.fn()
+}))
+
+vi.mock(import('../openai'), async (importOriginal) => {
+  const actual = await importOriginal()
+  return {
+    ...actual,
+    isOpenAIReasoningModel: vi.fn()
+  }
+})
+
+vi.mock('../websearch', () => ({
+  isOpenAIWebSearchChatCompletionOnlyModel: vi.fn()
+}))
+
+const createModel = (overrides: Partial<Model> = {}): Model => ({
+  id: 'gpt-4o',
+  name: 'gpt-4o',
+  provider: 'openai',
+  group: 'OpenAI',
+  ...overrides
+})
+
+const embeddingMock = vi.mocked(isEmbeddingModel)
+const rerankMock = vi.mocked(isRerankModel)
+const visionMock = vi.mocked(isVisionModel)
+const textToImageMock = vi.mocked(isTextToImageModel)
+const generateImageMock = vi.mocked(isGenerateImageModel)
+const reasoningMock = vi.mocked(isOpenAIReasoningModel)
+const openAIWebSearchOnlyMock = vi.mocked(isOpenAIWebSearchChatCompletionOnlyModel)
+
+describe('model utils', () => {
+  beforeEach(() => {
+    vi.clearAllMocks()
+    embeddingMock.mockReturnValue(false)
+    rerankMock.mockReturnValue(false)
+    visionMock.mockReturnValue(true)
+    textToImageMock.mockReturnValue(false)
+    generateImageMock.mockReturnValue(true)
+    reasoningMock.mockReturnValue(false)
+    openAIWebSearchOnlyMock.mockReturnValue(false)
+  })
+
+  it('detects OpenAI LLM models through reasoning and GPT prefix', () => {
+    expect(isOpenAILLMModel(undefined as unknown as Model)).toBe(false)
+    expect(isOpenAILLMModel(createModel({ id: 'gpt-4o-image' }))).toBe(false)
+
+    reasoningMock.mockReturnValueOnce(true)
+    expect(isOpenAILLMModel(createModel({ id: 'o1-preview' }))).toBe(true)
+
+    expect(isOpenAILLMModel(createModel({ id: 'GPT-5-turbo' }))).toBe(true)
+  })
+
+  it('detects OpenAI models via GPT prefix or reasoning support', () => {
+    expect(isOpenAIModel(createModel({ id: 'gpt-4.1' }))).toBe(true)
+    reasoningMock.mockReturnValueOnce(true)
+    expect(isOpenAIModel(createModel({ id: 'o3' }))).toBe(true)
+  })
+
+  it('evaluates support for flex service tier and alias helper', () => {
+    expect(isSupportFlexServiceTierModel(createModel({ id: 'o3' }))).toBe(true)
+    expect(isSupportFlexServiceTierModel(createModel({ id: 'o3-mini' }))).toBe(false)
+    expect(isSupportFlexServiceTierModel(createModel({ id: 'o4-mini' }))).toBe(true)
+    expect(isSupportFlexServiceTierModel(createModel({ id: 'gpt-5-preview' }))).toBe(true)
+    expect(isSupportedFlexServiceTier(createModel({ id: 'gpt-4o' }))).toBe(false)
+  })
+
+  it('detects verbosity support for GPT-5+ families', () => {
+    expect(isSupportVerbosityModel(createModel({ id: 'gpt-5' }))).toBe(true)
+    expect(isSupportVerbosityModel(createModel({ id: 'gpt-5-chat' }))).toBe(false)
+    expect(isSupportVerbosityModel(createModel({ id: 'gpt-5.1-preview' }))).toBe(true)
+  })
+
+  it('limits verbosity controls for GPT-5 Pro models', () => {
+    const proModel = createModel({ id: 'gpt-5-pro' })
+    const previewModel = createModel({ id: 'gpt-5-preview' })
+    expect(getModelSupportedVerbosity(proModel)).toEqual([undefined, 'high'])
+    expect(getModelSupportedVerbosity(previewModel)).toEqual([undefined, 'low', 'medium', 'high'])
+    expect(isGPT5ProModel(proModel)).toBe(true)
+    expect(isGPT5ProModel(previewModel)).toBe(false)
+  })
+
+  it('identifies OpenAI chat-completion-only models', () => {
+    expect(isOpenAIChatCompletionOnlyModel(createModel({ id: 'gpt-4o-search-preview' }))).toBe(true)
+    expect(isOpenAIChatCompletionOnlyModel(createModel({ id: 'o1-mini' }))).toBe(true)
+    expect(isOpenAIChatCompletionOnlyModel(createModel({ id: 'gpt-4o' }))).toBe(false)
+  })
+
+  it('filters unsupported OpenAI catalog entries', () => {
+    expect(isSupportedModel({ id: 'gpt-4', object: 'model' } as any)).toBe(true)
+    expect(isSupportedModel({ id: 'tts-1', object: 'model' } as any)).toBe(false)
+  })
+
+  it('calculates temperature/top-p support correctly', () => {
+    const model = createModel({ id: 'o1' })
+    reasoningMock.mockReturnValue(true)
+    expect(isNotSupportTemperatureAndTopP(model)).toBe(true)
+
+    const openWeight = createModel({ id: 'gpt-oss-debug' })
+    expect(isNotSupportTemperatureAndTopP(openWeight)).toBe(false)
+
+    const chatOnly = createModel({ id: 'o1-preview' })
+    reasoningMock.mockReturnValue(false)
+    expect(isNotSupportTemperatureAndTopP(chatOnly)).toBe(true)
+
+    const qwenMt = createModel({ id: 'qwen-mt-large', provider: 'aliyun' })
+    expect(isNotSupportTemperatureAndTopP(qwenMt)).toBe(true)
+  })
+
+  it('handles gemma and gemini detections plus zhipu tagging', () => {
+    expect(isGemmaModel(createModel({ id: 'Gemma-3-27B' }))).toBe(true)
+    expect(isGemmaModel(createModel({ group: 'Gemma' }))).toBe(true)
+    expect(isGemmaModel(createModel({ id: 'gpt-4o' }))).toBe(false)
+
+    expect(isGeminiModel(createModel({ id: 'Gemini-2.0' }))).toBe(true)
+
+    expect(isZhipuModel(createModel({ provider: 'zhipu' }))).toBe(true)
+    expect(isZhipuModel(createModel({ provider: 'openai' }))).toBe(false)
+  })
+
+  it('groups qwen models by prefix', () => {
+    const qwen = createModel({ id: 'Qwen-7B', provider: 'qwen', name: 'Qwen-7B' })
+    const qwenOmni = createModel({ id: 'qwen2.5-omni', name: 'qwen2.5-omni' })
+    const other = createModel({ id: 'deepseek-v3', group: 'DeepSeek' })
+
+    const grouped = groupQwenModels([qwen, qwenOmni, other])
+    expect(Object.keys(grouped)).toContain('qwen-7b')
+    expect(Object.keys(grouped)).toContain('qwen2.5')
+    expect(grouped.DeepSeek).toContain(other)
+  })
+
+  it('aggregates boolean helpers based on regex rules', () => {
+    expect(isAnthropicModel(createModel({ id: 'claude-3.5' }))).toBe(true)
+    expect(isQwenMTModel(createModel({ id: 'qwen-mt-large' }))).toBe(true)
+    expect(isNotSupportedTextDelta(createModel({ id: 'qwen-mt-large' }))).toBe(true)
+    expect(isNotSupportSystemMessageModel(createModel({ id: 'gemma-moe' }))).toBe(true)
+    expect(isOpenAIOpenWeightModel(createModel({ id: 'gpt-oss-free' }))).toBe(true)
+  })
+
+  it('evaluates GPT-5 family helpers', () => {
+    expect(isGPT5SeriesModel(createModel({ id: 'gpt-5-preview' }))).toBe(true)
+    expect(isGPT5SeriesModel(createModel({ id: 'gpt-5.1-preview' }))).toBe(false)
+    expect(isGPT51SeriesModel(createModel({ id: 'gpt-5.1-mini' }))).toBe(true)
+    expect(isGPT5SeriesReasoningModel(createModel({ id: 'gpt-5-prompt' }))).toBe(true)
+    expect(isSupportVerbosityModel(createModel({ id: 'gpt-5-chat' }))).toBe(false)
+  })
+
+  it('wraps generate/vision helpers that operate on arrays', () => {
+    const models = [createModel({ id: 'gpt-4o' }), createModel({ id: 'gpt-4o-mini' })]
+    expect(isVisionModels(models)).toBe(true)
+    visionMock.mockReturnValueOnce(true).mockReturnValueOnce(false)
+    expect(isVisionModels(models)).toBe(false)
+
+    expect(isGenerateImageModels(models)).toBe(true)
+    generateImageMock.mockReturnValueOnce(true).mockReturnValueOnce(false)
+    expect(isGenerateImageModels(models)).toBe(false)
+  })
+
+  it('filters models for agent usage', () => {
+    expect(agentModelFilter(createModel())).toBe(true)
+
+    embeddingMock.mockReturnValueOnce(true)
+    expect(agentModelFilter(createModel({ id: 'text-embedding' }))).toBe(false)
+
+    embeddingMock.mockReturnValue(false)
+    rerankMock.mockReturnValueOnce(true)
+    expect(agentModelFilter(createModel({ id: 'rerank' }))).toBe(false)
+
+    rerankMock.mockReturnValue(false)
+    textToImageMock.mockReturnValueOnce(true)
+    expect(agentModelFilter(createModel({ id: 'gpt-image-1' }))).toBe(false)
+  })
+
+  it('identifies models with maximum temperature of 1.0', () => {
+    // Zhipu models should have max temperature of 1.0
+    expect(isMaxTemperatureOneModel(createModel({ id: 'glm-4' }))).toBe(true)
+    expect(isMaxTemperatureOneModel(createModel({ id: 'GLM-4-Plus' }))).toBe(true)
+    expect(isMaxTemperatureOneModel(createModel({ id: 'glm-3-turbo' }))).toBe(true)
+
+    // Anthropic models should have max temperature of 1.0
+    expect(isMaxTemperatureOneModel(createModel({ id: 'claude-3.5-sonnet' }))).toBe(true)
+    expect(isMaxTemperatureOneModel(createModel({ id: 'Claude-3-opus' }))).toBe(true)
+    expect(isMaxTemperatureOneModel(createModel({ id: 'claude-2.1' }))).toBe(true)
+
+    // Moonshot models should have max temperature of 1.0
+    expect(isMaxTemperatureOneModel(createModel({ id: 'moonshot-1.0' }))).toBe(true)
+    expect(isMaxTemperatureOneModel(createModel({ id: 'kimi-k2-thinking' }))).toBe(true)
+    expect(isMaxTemperatureOneModel(createModel({ id: 'Moonshot-Pro' }))).toBe(true)
+
+    // Other models should return false
+    expect(isMaxTemperatureOneModel(createModel({ id: 'gpt-4o' }))).toBe(false)
+    expect(isMaxTemperatureOneModel(createModel({ id: 'gpt-4-turbo' }))).toBe(false)
+    expect(isMaxTemperatureOneModel(createModel({ id: 'qwen-max' }))).toBe(false)
+    expect(isMaxTemperatureOneModel(createModel({ id: 'gemini-pro' }))).toBe(false)
+  })
+})
@@ -0,0 +1,311 @@
+import { getProviderByModel } from '@renderer/services/AssistantService'
+import type { Model } from '@renderer/types'
+import { beforeEach, describe, expect, it, vi } from 'vitest'
+
+import { isEmbeddingModel, isRerankModel } from '../embedding'
+import {
+  isAutoEnableImageGenerationModel,
+  isDedicatedImageGenerationModel,
+  isGenerateImageModel,
+  isImageEnhancementModel,
+  isPureGenerateImageModel,
+  isTextToImageModel,
+  isVisionModel
+} from '../vision'
+
+vi.mock('@renderer/hooks/useStore', () => ({
+  getStoreProviders: vi.fn(() => [])
+}))
+
+vi.mock('@renderer/store', () => ({
+  __esModule: true,
+  default: {
+    getState: () => ({
+      llm: { providers: [] },
+      settings: {}
+    })
+  },
+  useAppDispatch: vi.fn(),
+  useAppSelector: vi.fn()
+}))
+
+vi.mock('@renderer/store/settings', () => {
+  const noop = vi.fn()
+  return new Proxy(
+    {},
+    {
+      get: (_target, prop) => {
+        if (prop === 'initialState') {
+          return {}
+        }
+        return noop
+      }
+    }
+  )
+})
+
+vi.mock('@renderer/hooks/useSettings', () => ({
+  useSettings: vi.fn(() => ({})),
+  useNavbarPosition: vi.fn(() => ({ navbarPosition: 'left' })),
+  useMessageStyle: vi.fn(() => ({ isBubbleStyle: false })),
+  getStoreSetting: vi.fn()
+}))
+
+vi.mock('@renderer/services/AssistantService', () => ({
+  getProviderByModel: vi.fn()
+}))
+
+vi.mock('../embedding', () => ({
+  isEmbeddingModel: vi.fn(),
+  isRerankModel: vi.fn()
+}))
+
+const createModel = (overrides: Partial<Model> = {}): Model => ({
+  id: 'gpt-4o',
+  name: 'gpt-4o',
+  provider: 'openai',
+  group: 'OpenAI',
+  ...overrides
+})
+
+const providerMock = vi.mocked(getProviderByModel)
+const embeddingMock = vi.mocked(isEmbeddingModel)
+const rerankMock = vi.mocked(isRerankModel)
+
+describe('vision helpers', () => {
+  beforeEach(() => {
+    vi.clearAllMocks()
+    providerMock.mockReturnValue({ type: 'openai-response' } as any)
+    embeddingMock.mockReturnValue(false)
+    rerankMock.mockReturnValue(false)
+  })
+
+  describe('isGenerateImageModel', () => {
+    it('returns false for embedding/rerank models or missing providers', () => {
+      embeddingMock.mockReturnValueOnce(true)
+      expect(isGenerateImageModel(createModel({ id: 'gpt-image-1' }))).toBe(false)
+
+      embeddingMock.mockReturnValue(false)
+      rerankMock.mockReturnValueOnce(true)
+      expect(isGenerateImageModel(createModel({ id: 'gpt-image-1' }))).toBe(false)
+
+      rerankMock.mockReturnValue(false)
+      providerMock.mockReturnValueOnce(undefined as any)
+      expect(isGenerateImageModel(createModel({ id: 'gpt-image-1' }))).toBe(false)
+    })
+
+    it('detects OpenAI and third-party generative image models', () => {
+      expect(isGenerateImageModel(createModel({ id: 'gpt-4o-mini' }))).toBe(true)
+
+      providerMock.mockReturnValue({ type: 'custom' } as any)
+      expect(isGenerateImageModel(createModel({ id: 'gemini-2.5-flash-image' }))).toBe(true)
+    })
+
+    it('returns false when openai-response model is not on allow list', () => {
+      expect(isGenerateImageModel(createModel({ id: 'gpt-4.2-experimental' }))).toBe(false)
+    })
+  })
+
+  describe('isPureGenerateImageModel', () => {
+    it('requires both generate and text-to-image support', () => {
+      expect(isPureGenerateImageModel(createModel({ id: 'gpt-image-1' }))).toBe(true)
+      expect(isPureGenerateImageModel(createModel({ id: 'gpt-4o' }))).toBe(false)
+      expect(isPureGenerateImageModel(createModel({ id: 'gemini-2.5-flash-image-preview' }))).toBe(true)
+    })
+  })
+
+  describe('text-to-image helpers', () => {
+    it('matches predefined keywords', () => {
+      expect(isTextToImageModel(createModel({ id: 'midjourney-v6' }))).toBe(true)
+      expect(isTextToImageModel(createModel({ id: 'gpt-4o' }))).toBe(false)
+    })
+
+    it('detects models with restricted image size support and enhancement', () => {
+      expect(isImageEnhancementModel(createModel({ id: 'qwen-image-edit' }))).toBe(true)
+      expect(isImageEnhancementModel(createModel({ id: 'gpt-4o' }))).toBe(false)
+    })
+
+    it('identifies dedicated and auto-enabled image generation models', () => {
+      expect(isDedicatedImageGenerationModel(createModel({ id: 'grok-2-image-1212' }))).toBe(true)
+      expect(isAutoEnableImageGenerationModel(createModel({ id: 'gemini-2.5-flash-image-ultra' }))).toBe(true)
+    })
+
+    it('returns false when models are not in dedicated or auto-enable sets', () => {
+      expect(isDedicatedImageGenerationModel(createModel({ id: 'gpt-4o' }))).toBe(false)
+      expect(isAutoEnableImageGenerationModel(createModel({ id: 'gpt-4o' }))).toBe(false)
+    })
+  })
+})
+
+describe('isVisionModel', () => {
+  it('returns false for embedding/rerank models and honors overrides', () => {
+    embeddingMock.mockReturnValueOnce(true)
+    expect(isVisionModel(createModel({ id: 'gpt-4o' }))).toBe(false)
+
+    embeddingMock.mockReturnValue(false)
+    const disabled = createModel({
+      id: 'gpt-4o',
+      capabilities: [{ type: 'vision', isUserSelected: false }]
+    })
+    expect(isVisionModel(disabled)).toBe(false)
+
+    const forced = createModel({
+      id: 'gpt-4o',
+      capabilities: [{ type: 'vision', isUserSelected: true }]
+    })
+    expect(isVisionModel(forced)).toBe(true)
+  })
+
+  it('matches doubao models by name and general regexes by id', () => {
+    const doubao = createModel({
+      id: 'custom-id',
+      provider: 'doubao',
+      name: 'Doubao-Seed-1-6-Lite-251015'
+    })
+    expect(isVisionModel(doubao)).toBe(true)
+
+    expect(isVisionModel(createModel({ id: 'gpt-4o-mini' }))).toBe(true)
+  })
+
+  it('leverages image enhancement regex when standard vision regex does not match', () => {
+    expect(isVisionModel(createModel({ id: 'qwen-image-edit' }))).toBe(true)
+  })
+
+  it('returns false for doubao models that fail regex checks', () => {
+    const doubao = createModel({ id: 'doubao-standard', provider: 'doubao', name: 'basic' })
+    expect(isVisionModel(doubao)).toBe(false)
+  })
+  describe('Gemini Models', () => {
+    it('should return true for gemini 1.5 models', () => {
+      expect(
+        isVisionModel({
+          id: 'gemini-1.5-flash',
+          name: '',
+          provider: '',
+          group: ''
+        })
+      ).toBe(true)
+      expect(
+        isVisionModel({
+          id: 'gemini-1.5-pro',
+          name: '',
+          provider: '',
+          group: ''
+        })
+      ).toBe(true)
+    })
+
+    it('should return true for gemini 2.x models', () => {
+      expect(
+        isVisionModel({
+          id: 'gemini-2.0-flash',
+          name: '',
+          provider: '',
+          group: ''
+        })
+      ).toBe(true)
+      expect(
+        isVisionModel({
+          id: 'gemini-2.0-pro',
+          name: '',
+          provider: '',
+          group: ''
+        })
+      ).toBe(true)
+      expect(
+        isVisionModel({
+          id: 'gemini-2.5-flash',
+          name: '',
+          provider: '',
+          group: ''
+        })
+      ).toBe(true)
+      expect(
+        isVisionModel({
+          id: 'gemini-2.5-pro',
+          name: '',
+          provider: '',
+          group: ''
+        })
+      ).toBe(true)
+    })
+
+    it('should return true for gemini latest models', () => {
+      expect(
+        isVisionModel({
+          id: 'gemini-flash-latest',
+          name: '',
+          provider: '',
+          group: ''
+        })
+      ).toBe(true)
+      expect(
+        isVisionModel({
+          id: 'gemini-pro-latest',
+          name: '',
+          provider: '',
+          group: ''
+        })
+      ).toBe(true)
+      expect(
+        isVisionModel({
+          id: 'gemini-flash-lite-latest',
+          name: '',
+          provider: '',
+          group: ''
+        })
+      ).toBe(true)
+    })
+
+    it('should return true for gemini 3 models', () => {
+      // Preview versions
+      expect(
+        isVisionModel({
+          id: 'gemini-3-pro-preview',
+          name: '',
+          provider: '',
+          group: ''
+        })
+      ).toBe(true)
+      // Future stable versions
+      expect(
+        isVisionModel({
+          id: 'gemini-3-flash',
+          name: '',
+          provider: '',
+          group: ''
+        })
+      ).toBe(true)
+      expect(
+        isVisionModel({
+          id: 'gemini-3-pro',
+          name: '',
+          provider: '',
+          group: ''
+        })
+      ).toBe(true)
+    })
+
+    it('should return true for gemini exp models', () => {
+      expect(
+        isVisionModel({
+          id: 'gemini-exp-1206',
+          name: '',
+          provider: '',
+          group: ''
+        })
+      ).toBe(true)
+    })
+
+    it('should return false for gemini 1.0 models', () => {
+      expect(
+        isVisionModel({
+          id: 'gemini-1.0-pro',
+          name: '',
+          provider: '',
+          group: ''
+        })
+      ).toBe(false)
+    })
+  })
+})
@@ -0,0 +1,397 @@
+import { beforeEach, describe, expect, it, vi } from 'vitest'
+
+const providerMock = vi.mocked(getProviderByModel)
+
+vi.mock('@renderer/services/AssistantService', () => ({
+  getProviderByModel: vi.fn(),
+  getAssistantSettings: vi.fn(),
+  getDefaultAssistant: vi.fn().mockReturnValue({
+    id: 'default',
+    name: 'Default Assistant',
+    prompt: '',
+    settings: {}
+  })
+}))
+
+const isEmbeddingModel = vi.hoisted(() => vi.fn())
+const isRerankModel = vi.hoisted(() => vi.fn())
+vi.mock('../embedding', () => ({
+  isEmbeddingModel: (...args: any[]) => isEmbeddingModel(...args),
+  isRerankModel: (...args: any[]) => isRerankModel(...args)
+}))
+
+const isPureGenerateImageModel = vi.hoisted(() => vi.fn())
+const isTextToImageModel = vi.hoisted(() => vi.fn())
+const isGenerateImageModel = vi.hoisted(() => vi.fn())
+vi.mock('../vision', () => ({
+  isPureGenerateImageModel: (...args: any[]) => isPureGenerateImageModel(...args),
+  isTextToImageModel: (...args: any[]) => isTextToImageModel(...args),
+  isGenerateImageModel: (...args: any[]) => isGenerateImageModel(...args),
+  isModernGenerateImageModel: vi.fn()
+}))
+
+const providerMocks = vi.hoisted(() => ({
+  isGeminiProvider: vi.fn(),
+  isNewApiProvider: vi.fn(),
+  isOpenAICompatibleProvider: vi.fn(),
+  isOpenAIProvider: vi.fn(),
+  isVertexProvider: vi.fn(),
+  isAwsBedrockProvider: vi.fn(),
+  isAzureOpenAIProvider: vi.fn()
+}))
+
+vi.mock('@renderer/utils/provider', () => providerMocks)
+
+vi.mock('@renderer/hooks/useStore', () => ({
+  getStoreProviders: vi.fn(() => [])
+}))
+
+vi.mock('@renderer/store', () => ({
+  __esModule: true,
+  default: {
+    getState: () => ({
+      llm: { providers: [] },
+      settings: {}
+    })
+  },
+  useAppDispatch: vi.fn(),
+  useAppSelector: vi.fn()
+}))
+
+vi.mock('@renderer/store/settings', () => {
+  const noop = vi.fn()
+  return new Proxy(
+    {},
+    {
+      get: (_target, prop) => {
+        if (prop === 'initialState') {
+          return {}
+        }
+        return noop
+      }
+    }
+  )
+})
+
+vi.mock('@renderer/hooks/useSettings', () => ({
+  useSettings: vi.fn(() => ({})),
+  useNavbarPosition: vi.fn(() => ({ navbarPosition: 'left' })),
+  useMessageStyle: vi.fn(() => ({ isBubbleStyle: false })),
+  getStoreSetting: vi.fn()
+}))
+
+import { getProviderByModel } from '@renderer/services/AssistantService'
+import type { Model, Provider } from '@renderer/types'
+import { SystemProviderIds } from '@renderer/types'
+
+import { isOpenAIDeepResearchModel } from '../openai'
+import {
+  GEMINI_SEARCH_REGEX,
+  isHunyuanSearchModel,
+  isMandatoryWebSearchModel,
+  isOpenAIWebSearchChatCompletionOnlyModel,
+  isOpenAIWebSearchModel,
+  isOpenRouterBuiltInWebSearchModel,
+  isWebSearchModel
+} from '../websearch'
+
+const createModel = (overrides: Partial<Model> = {}): Model => ({
+  id: 'gpt-4o',
+  name: 'gpt-4o',
+  provider: 'openai',
+  group: 'OpenAI',
+  ...overrides
+})
+
+const createProvider = (overrides: Partial<Provider> = {}): Provider => ({
+  id: 'openai',
+  type: 'openai',
+  name: 'OpenAI',
+  apiKey: '',
+  apiHost: '',
+  models: [],
+  ...overrides
+})
+
+const resetMocks = () => {
+  providerMock.mockReturnValue(createProvider())
+  isEmbeddingModel.mockReturnValue(false)
+  isRerankModel.mockReturnValue(false)
+  isPureGenerateImageModel.mockReturnValue(false)
+  isTextToImageModel.mockReturnValue(false)
+  providerMocks.isGeminiProvider.mockReturnValue(false)
+  providerMocks.isNewApiProvider.mockReturnValue(false)
+  providerMocks.isOpenAICompatibleProvider.mockReturnValue(false)
+  providerMocks.isOpenAIProvider.mockReturnValue(false)
+}
+
+describe('websearch helpers', () => {
+  beforeEach(() => {
+    vi.clearAllMocks()
+    resetMocks()
+  })
+
+  describe('isOpenAIDeepResearchModel', () => {
+    it('detects deep research ids for OpenAI only', () => {
+      expect(isOpenAIDeepResearchModel(createModel({ id: 'openai/deep-research-preview' }))).toBe(true)
+      expect(isOpenAIDeepResearchModel(createModel({ provider: 'openai', id: 'gpt-4o' }))).toBe(false)
+      expect(isOpenAIDeepResearchModel(createModel({ provider: 'openrouter', id: 'deep-research' }))).toBe(false)
+    })
+  })
+
+  describe('isWebSearchModel', () => {
+    it('returns false for embedding/rerank/image models', () => {
+      isEmbeddingModel.mockReturnValueOnce(true)
+      expect(isWebSearchModel(createModel())).toBe(false)
+
+      resetMocks()
+      isRerankModel.mockReturnValueOnce(true)
+      expect(isWebSearchModel(createModel())).toBe(false)
+
+      resetMocks()
+      isTextToImageModel.mockReturnValueOnce(true)
+      expect(isWebSearchModel(createModel())).toBe(false)
+    })
+
+    it('honors user overrides', () => {
+      const enabled = createModel({ capabilities: [{ type: 'web_search', isUserSelected: true }] })
+      expect(isWebSearchModel(enabled)).toBe(true)
+
+      const disabled = createModel({ capabilities: [{ type: 'web_search', isUserSelected: false }] })
+      expect(isWebSearchModel(disabled)).toBe(false)
+    })
+
+    it('returns false when provider lookup fails', () => {
+      providerMock.mockReturnValueOnce(undefined as any)
+      expect(isWebSearchModel(createModel())).toBe(false)
+    })
+
+    it('handles Anthropic providers on unsupported platforms', () => {
+      providerMock.mockReturnValueOnce(createProvider({ id: SystemProviderIds['aws-bedrock'] }))
+      const model = createModel({ id: 'claude-2-sonnet' })
+      expect(isWebSearchModel(model)).toBe(false)
+    })
+
+    it('returns true for first-party Anthropic provider', () => {
+      providerMock.mockReturnValueOnce(createProvider({ id: 'anthropic' }))
+      const model = createModel({ id: 'claude-3.5-sonnet-latest', provider: 'anthropic' })
+      expect(isWebSearchModel(model)).toBe(true)
+    })
+
+    it('detects OpenAI preview search models only when supported', () => {
+      providerMocks.isOpenAIProvider.mockReturnValue(true)
+      const model = createModel({ id: 'gpt-4o-search-preview' })
+      expect(isWebSearchModel(model)).toBe(true)
+
+      const nonSearch = createModel({ id: 'gpt-4o-image' })
+      expect(isWebSearchModel(nonSearch)).toBe(false)
+    })
+
+    it('supports Perplexity sonar families including mandatory variants', () => {
+      providerMock.mockReturnValueOnce(createProvider({ id: SystemProviderIds.perplexity }))
+      expect(isWebSearchModel(createModel({ id: 'sonar-deep-research' }))).toBe(true)
+    })
+
+    it('handles AIHubMix Gemini and OpenAI search models', () => {
+      providerMock.mockReturnValueOnce(createProvider({ id: SystemProviderIds.aihubmix }))
+      expect(isWebSearchModel(createModel({ id: 'gemini-2.5-pro-preview' }))).toBe(true)
+
+      providerMock.mockReturnValueOnce(createProvider({ id: SystemProviderIds.aihubmix }))
+      const openaiSearch = createModel({ id: 'gpt-4o-search-preview' })
+      expect(isWebSearchModel(openaiSearch)).toBe(true)
+    })
+
+    it('supports OpenAI-compatible or new API providers for Gemini/OpenAI models', () => {
+      const model = createModel({ id: 'gemini-2.5-flash-lite-latest' })
+      providerMock.mockReturnValueOnce(createProvider({ id: 'custom' }))
+      providerMocks.isOpenAICompatibleProvider.mockReturnValueOnce(true)
+      expect(isWebSearchModel(model)).toBe(true)
+
+      resetMocks()
+      providerMock.mockReturnValueOnce(createProvider({ id: 'custom' }))
+      providerMocks.isNewApiProvider.mockReturnValueOnce(true)
+      expect(isWebSearchModel(createModel({ id: 'gpt-4o-search-preview' }))).toBe(true)
+    })
+
+    it('falls back to Gemini/Vertex provider regex matching', () => {
+      providerMock.mockReturnValueOnce(createProvider({ id: SystemProviderIds.vertexai }))
+      providerMocks.isGeminiProvider.mockReturnValueOnce(true)
+      expect(isWebSearchModel(createModel({ id: 'gemini-2.0-flash-latest' }))).toBe(true)
+    })
+
+    it('evaluates hunyuan/zhipu/dashscope/openrouter/grok providers', () => {
+      providerMock.mockReturnValueOnce(createProvider({ id: 'hunyuan' }))
+      expect(isWebSearchModel(createModel({ id: 'hunyuan-pro' }))).toBe(true)
+      expect(isWebSearchModel(createModel({ id: 'hunyuan-lite', provider: 'hunyuan' }))).toBe(false)
+
+      providerMock.mockReturnValueOnce(createProvider({ id: 'zhipu' }))
+      expect(isWebSearchModel(createModel({ id: 'glm-4-air' }))).toBe(true)
+
+      providerMock.mockReturnValueOnce(createProvider({ id: 'dashscope' }))
+      expect(isWebSearchModel(createModel({ id: 'qwen-max-latest' }))).toBe(true)
+
+      providerMock.mockReturnValueOnce(createProvider({ id: 'openrouter' }))
+      expect(isWebSearchModel(createModel())).toBe(true)
+
+      providerMock.mockReturnValueOnce(createProvider({ id: 'grok' }))
+      expect(isWebSearchModel(createModel({ id: 'grok-2' }))).toBe(true)
+    })
+  })
+
+  describe('isMandatoryWebSearchModel', () => {
+    it('requires sonar ids for perplexity/openrouter providers', () => {
+      providerMock.mockReturnValueOnce(createProvider({ id: SystemProviderIds.perplexity }))
+      expect(isMandatoryWebSearchModel(createModel({ id: 'sonar-pro' }))).toBe(true)
+
+      providerMock.mockReturnValueOnce(createProvider({ id: SystemProviderIds.openrouter }))
+      expect(isMandatoryWebSearchModel(createModel({ id: 'sonar-reasoning' }))).toBe(true)
+
+      providerMock.mockReturnValueOnce(createProvider({ id: 'openai' }))
+      expect(isMandatoryWebSearchModel(createModel({ id: 'sonar-pro' }))).toBe(false)
+    })
+
+    it.each([
+      ['perplexity', 'non-sonar'],
+      ['openrouter', 'gpt-4o-search-preview']
+    ])('returns false for %s provider when id is %s', (providerId, modelId) => {
+      providerMock.mockReturnValueOnce(createProvider({ id: providerId }))
+      expect(isMandatoryWebSearchModel(createModel({ id: modelId }))).toBe(false)
+    })
+  })
+
+  describe('isOpenRouterBuiltInWebSearchModel', () => {
+    it('checks for sonar ids or OpenAI chat-completion-only variants', () => {
+      providerMock.mockReturnValueOnce(createProvider({ id: 'openrouter' }))
+      expect(isOpenRouterBuiltInWebSearchModel(createModel({ id: 'sonar-reasoning' }))).toBe(true)
+
+      providerMock.mockReturnValueOnce(createProvider({ id: 'openrouter' }))
+      expect(isOpenRouterBuiltInWebSearchModel(createModel({ id: 'gpt-4o-search-preview' }))).toBe(true)
+
+      providerMock.mockReturnValueOnce(createProvider({ id: 'custom' }))
+      expect(isOpenRouterBuiltInWebSearchModel(createModel({ id: 'sonar-reasoning' }))).toBe(false)
+    })
+  })
+
+  describe('OpenAI web search helpers', () => {
+    it('detects chat completion only variants and openai search ids', () => {
+      expect(isOpenAIWebSearchChatCompletionOnlyModel(createModel({ id: 'gpt-4o-search-preview' }))).toBe(true)
+      expect(isOpenAIWebSearchChatCompletionOnlyModel(createModel({ id: 'gpt-4o-mini-search-preview' }))).toBe(true)
+      expect(isOpenAIWebSearchChatCompletionOnlyModel(createModel({ id: 'gpt-4o' }))).toBe(false)
+
+      expect(isOpenAIWebSearchModel(createModel({ id: 'gpt-4.1-turbo' }))).toBe(true)
+      expect(isOpenAIWebSearchModel(createModel({ id: 'gpt-4o-image' }))).toBe(false)
+      expect(isOpenAIWebSearchModel(createModel({ id: 'gpt-5.1-chat' }))).toBe(false)
+      expect(isOpenAIWebSearchModel(createModel({ id: 'o3-mini' }))).toBe(true)
+    })
+
+    it.each(['gpt-4.1-preview', 'gpt-4o-2024-05-13', 'o4-mini', 'gpt-5-explorer'])(
+      'treats %s as an OpenAI web search model',
+      (id) => {
+        expect(isOpenAIWebSearchModel(createModel({ id }))).toBe(true)
+      }
+    )
+
+    it.each(['gpt-4o-image-preview', 'gpt-4.1-nano', 'gpt-5.1-chat', 'gpt-image-1'])(
+      'excludes %s from OpenAI web search',
+      (id) => {
+        expect(isOpenAIWebSearchModel(createModel({ id }))).toBe(false)
+      }
+    )
+
+    it.each(['gpt-4o-search-preview', 'gpt-4o-mini-search-preview'])('flags %s as chat-completion-only', (id) => {
+      expect(isOpenAIWebSearchChatCompletionOnlyModel(createModel({ id }))).toBe(true)
+    })
+  })
+
+  describe('isHunyuanSearchModel', () => {
+    it('identifies hunyuan models except lite', () => {
+      expect(isHunyuanSearchModel(createModel({ id: 'hunyuan-pro', provider: 'hunyuan' }))).toBe(true)
+      expect(isHunyuanSearchModel(createModel({ id: 'hunyuan-lite', provider: 'hunyuan' }))).toBe(false)
+      expect(isHunyuanSearchModel(createModel())).toBe(false)
+    })
+
+    it.each(['hunyuan-standard', 'hunyuan-advanced'])('accepts %s', (suffix) => {
+      expect(isHunyuanSearchModel(createModel({ id: suffix, provider: 'hunyuan' }))).toBe(true)
+    })
+  })
+
+  describe('provider-specific regex coverage', () => {
+    it.each(['qwen-turbo', 'qwen-max-0919', 'qwen3-max', 'qwen-plus-2024', 'qwq-32b'])(
+      'dashscope treats %s as searchable',
+      (id) => {
+        providerMock.mockReturnValue(createProvider({ id: 'dashscope' }))
+        expect(isWebSearchModel(createModel({ id }))).toBe(true)
+      }
+    )
+
+    it.each(['qwen-1.5-chat', 'custom-model'])('dashscope ignores %s', (id) => {
+      providerMock.mockReturnValue(createProvider({ id: 'dashscope' }))
+      expect(isWebSearchModel(createModel({ id }))).toBe(false)
+    })
+
+    it.each(['sonar', 'sonar-pro', 'sonar-reasoning-pro', 'sonar-deep-research'])(
+      'perplexity provider supports %s',
+      (id) => {
+        providerMock.mockReturnValue(createProvider({ id: SystemProviderIds.perplexity }))
+        expect(isWebSearchModel(createModel({ id }))).toBe(true)
+      }
+    )
+
+    it.each([
+      'gemini-2.0-flash-latest',
+      'gemini-2.5-flash-lite-latest',
+      'gemini-flash-lite-latest',
+      'gemini-pro-latest'
+    ])('Gemini provider supports %s', (id) => {
+      providerMock.mockReturnValue(createProvider({ id: SystemProviderIds.vertexai }))
+      providerMocks.isGeminiProvider.mockReturnValue(true)
+      expect(isWebSearchModel(createModel({ id }))).toBe(true)
+    })
+  })
+
+  describe('Gemini Search Models', () => {
+    describe('GEMINI_SEARCH_REGEX', () => {
+      it('should match gemini 2.x models', () => {
+        expect(GEMINI_SEARCH_REGEX.test('gemini-2.0-flash')).toBe(true)
+        expect(GEMINI_SEARCH_REGEX.test('gemini-2.0-pro')).toBe(true)
+        expect(GEMINI_SEARCH_REGEX.test('gemini-2.5-flash')).toBe(true)
+        expect(GEMINI_SEARCH_REGEX.test('gemini-2.5-pro')).toBe(true)
+        expect(GEMINI_SEARCH_REGEX.test('gemini-2.5-flash-latest')).toBe(true)
+        expect(GEMINI_SEARCH_REGEX.test('gemini-2.5-pro-latest')).toBe(true)
+      })
+
+      it('should match gemini latest models', () => {
+        expect(GEMINI_SEARCH_REGEX.test('gemini-flash-latest')).toBe(true)
+        expect(GEMINI_SEARCH_REGEX.test('gemini-pro-latest')).toBe(true)
+        expect(GEMINI_SEARCH_REGEX.test('gemini-flash-lite-latest')).toBe(true)
+      })
+
+      it('should match gemini 3 models', () => {
+        // Preview versions
+        expect(GEMINI_SEARCH_REGEX.test('gemini-3-pro-preview')).toBe(true)
+        expect(GEMINI_SEARCH_REGEX.test('gemini-3-flash-preview')).toBe(true)
+        expect(GEMINI_SEARCH_REGEX.test('gemini-3-pro-image-preview')).toBe(true)
+        expect(GEMINI_SEARCH_REGEX.test('gemini-3-flash-image-preview')).toBe(true)
+        // Future stable versions
+        expect(GEMINI_SEARCH_REGEX.test('gemini-3-flash')).toBe(true)
+        expect(GEMINI_SEARCH_REGEX.test('gemini-3-pro')).toBe(true)
+        // Version with decimals
+        expect(GEMINI_SEARCH_REGEX.test('gemini-3.0-flash')).toBe(true)
+        expect(GEMINI_SEARCH_REGEX.test('gemini-3.0-pro')).toBe(true)
+        expect(GEMINI_SEARCH_REGEX.test('gemini-3.5-flash-preview')).toBe(true)
+        expect(GEMINI_SEARCH_REGEX.test('gemini-3.5-pro-image-preview')).toBe(true)
+      })
+
+      it('should not match gemini 2.x image-preview models', () => {
+        expect(GEMINI_SEARCH_REGEX.test('gemini-2.5-flash-image-preview')).toBe(false)
+        expect(GEMINI_SEARCH_REGEX.test('gemini-2.0-pro-image-preview')).toBe(false)
+      })
+
+      it('should not match older gemini models', () => {
+        expect(GEMINI_SEARCH_REGEX.test('gemini-1.5-flash')).toBe(false)
+        expect(GEMINI_SEARCH_REGEX.test('gemini-1.5-pro')).toBe(false)
+        expect(GEMINI_SEARCH_REGEX.test('gemini-1.0-pro')).toBe(false)
+      })
+    })
+  })
+})
@@ -1,6 +1,8 @@
 export * from './default'
 export * from './embedding'
 export * from './logo'
+export * from './openai'
+export * from './qwen'
 export * from './reasoning'
 export * from './tooluse'
 export * from './utils'
@@ -59,6 +59,10 @@ import {
 } from '@renderer/assets/images/models/gpt_dark.png'
 import ChatGPTImageModelLogo from '@renderer/assets/images/models/gpt_image_1.png'
 import ChatGPTo1ModelLogo from '@renderer/assets/images/models/gpt_o1.png'
+import GPT51ModelLogo from '@renderer/assets/images/models/gpt-5.1.png'
+import GPT51ChatModelLogo from '@renderer/assets/images/models/gpt-5.1-chat.png'
+import GPT51CodexModelLogo from '@renderer/assets/images/models/gpt-5.1-codex.png'
+import GPT51CodexMiniModelLogo from '@renderer/assets/images/models/gpt-5.1-codex-mini.png'
 import GPT5ModelLogo from '@renderer/assets/images/models/gpt-5.png'
 import GPT5ChatModelLogo from '@renderer/assets/images/models/gpt-5-chat.png'
 import GPT5CodexModelLogo from '@renderer/assets/images/models/gpt-5-codex.png'
@@ -182,6 +186,10 @@ export function getModelLogoById(modelId: string): string | undefined {
    'gpt-5-nano': GPT5NanoModelLogo,
    'gpt-5-chat': GPT5ChatModelLogo,
    'gpt-5-codex': GPT5CodexModelLogo,
+    'gpt-5.1-codex-mini': GPT51CodexMiniModelLogo,
+    'gpt-5.1-codex': GPT51CodexModelLogo,
+    'gpt-5.1-chat': GPT51ChatModelLogo,
+    'gpt-5.1': GPT51ModelLogo,
    'gpt-5': GPT5ModelLogo,
    gpts: isLight ? ChatGPT4ModelLogo : ChatGPT4ModelLogoDark,
    'gpt-oss(?:-[\\w-]+)': isLight ? ChatGptModelLogo : ChatGptModelLogoDark,
--- a/Show More
+++ b/Show More
				`@@ -0,0 +1 @@`
				`export { default as MCPUIRenderer } from './MCPUIRenderer'`