Add initial chunk emission for agent response processing

- Emit `ChunkType.LLM_RESPONSE_CREATED` event to mirror assistant behavior
- Ensure UI reflects pending state immediately during response processing
- Maintain consistency with existing stream processing callback structure

.github/CODEOWNERS

@@ -0,0 +1,4 @@
+/src/renderer/src/store/ @0xfullex
+/src/main/services/ConfigManager.ts @0xfullex
+/packages/shared/IpcChannel.ts @0xfullex
+/src/main/ipc.ts @0xfullex

.github/ISSUE_TEMPLATE/#0_bug_report.yml

@@ -1,94 +0,0 @@
-name: 🐛 错误报告 (中文)
-description: 创建一个报告以帮助我们改进
-title: '[错误]: '
-labels: ['BUG']
-body:
-  - type: markdown
-    attributes:
-      value: |
-        感谢您花时间填写此错误报告！
-        在提交此问题之前，请确保您已经了解了[常见问题](https://docs.cherry-ai.com/question-contact/questions)和[知识科普](https://docs.cherry-ai.com/question-contact/knowledge)
-
-  - type: checkboxes
-    id: checklist
-    attributes:
-      label: 提交前检查
-      description: |
-        在提交 Issue 前请确保您已经完成了以下所有步骤
-      options:
-        - label: 我理解 Issue 是用于反馈和解决问题的，而非吐槽评论区，将尽可能提供更多信息帮助问题解决。
-          required: true
-        - label: 我的问题不是 [常见问题](https://github.com/CherryHQ/cherry-studio/issues/3881) 中的内容。
-          required: true
-        - label: 我已经查看了 **置顶 Issue** 并搜索了现有的 [开放Issue](https://github.com/CherryHQ/cherry-studio/issues)和[已关闭Issue](https://github.com/CherryHQ/cherry-studio/issues?q=is%3Aissue%20state%3Aclosed%20)，没有找到类似的问题。
-          required: true
-        - label: 我填写了简短且清晰明确的标题，以便开发者在翻阅 Issue 列表时能快速确定大致问题。而不是“一个建议”、“卡住了”等。
-          required: true
-        - label: 我确认我正在使用最新版本的 Cherry Studio。
-          required: true
-
-  - type: dropdown
-    id: platform
-    attributes:
-      label: 平台
-      description: 您正在使用哪个平台？
-      options:
-        - Windows
-        - macOS
-        - Linux
-    validations:
-      required: true
-
-  - type: input
-    id: version
-    attributes:
-      label: 版本
-      description: 您正在运行的 Cherry Studio 版本是什么？
-      placeholder: 例如 v1.0.0
-    validations:
-      required: true
-
-  - type: textarea
-    id: description
-    attributes:
-      label: 错误描述
-      description: 描述问题时请尽可能详细。请尽可能提供截图或屏幕录制，以帮助我们更好地理解问题。
-      placeholder: 告诉我们发生了什么...(记得附上截图/录屏，如果适用)
-    validations:
-      required: true
-
-  - type: textarea
-    id: reproduction
-    attributes:
-      label: 重现步骤
-      description: 提供详细的重现步骤，以便于我们的开发人员可以准确地重现问题。请尽可能为每个步骤提供截图或屏幕录制。
-      placeholder: |
-        1. 转到 '...'
-        2. 点击 '....'
-        3. 向下滚动到 '....'
-        4. 看到错误
-
-        记得尽可能为每个步骤附上截图/录屏！
-    validations:
-      required: true
-
-  - type: textarea
-    id: expected
-    attributes:
-      label: 预期行为
-      description: 清晰简洁地描述您期望发生的事情
-    validations:
-      required: true
-
-  - type: textarea
-    id: logs
-    attributes:
-      label: 相关日志输出
-      description: 请复制并粘贴任何相关的日志输出
-      render: shell
-
-  - type: textarea
-    id: additional
-    attributes:
-      label: 附加信息
-      description: 任何能让我们对你所遇到的问题有更多了解的东西

.github/ISSUE_TEMPLATE/#1_feature_request.yml

@@ -1,76 +0,0 @@
-name: 💡 功能建议 (中文)
-description: 为项目提出新的想法
-title: '[功能]: '
-labels: ['feature']
-body:
-  - type: markdown
-    attributes:
-      value: |
-        感谢您花时间提出新的功能建议!
-        在提交此问题之前，请确保您已经了解了[项目规划](https://docs.cherry-ai.com/cherrystudio/planning)和[功能介绍](https://docs.cherry-ai.com/cherrystudio/preview)
-
-  - type: checkboxes
-    id: checklist
-    attributes:
-      label: 提交前检查
-      description: |
-        在提交 Issue 前请确保您已经完成了以下所有步骤
-      options:
-        - label: 我理解 Issue 是用于反馈和解决问题的，而非吐槽评论区，将尽可能提供更多信息帮助问题解决。
-          required: true
-        - label: 我已经查看了置顶 Issue 并搜索了现有的 [开放Issue](https://github.com/CherryHQ/cherry-studio/issues)和[已关闭Issue](https://github.com/CherryHQ/cherry-studio/issues?q=is%3Aissue%20state%3Aclosed%20)，没有找到类似的建议。
-          required: true
-        - label: 我填写了简短且清晰明确的标题，以便开发者在翻阅 Issue 列表时能快速确定大致问题。而不是“一个建议”、“卡住了”等。
-          required: true
-        - label: 最新的 Cherry Studio 版本没有实现我所提出的功能。
-          required: true
-
-  - type: dropdown
-    id: platform
-    attributes:
-      label: 平台
-      description: 您正在使用哪个平台？
-      options:
-        - Windows
-        - macOS
-        - Linux
-    validations:
-      required: true
-
-  - type: input
-    id: version
-    attributes:
-      label: 版本
-      description: 您正在运行的 Cherry Studio 版本是什么？
-      placeholder: 例如 v1.0.0
-    validations:
-      required: true
-
-  - type: textarea
-    id: problem
-    attributes:
-      label: 您的功能建议是否与某个问题/issue相关?
-      description: 请简明扼要地描述您遇到的问题
-      placeholder: 我总是感到沮丧,因为...
-    validations:
-      required: true
-
-  - type: textarea
-    id: solution
-    attributes:
-      label: 请描述您希望实现的解决方案
-      description: 请简明扼要地描述您希望发生的情况
-    validations:
-      required: true
-
-  - type: textarea
-    id: alternatives
-    attributes:
-      label: 请描述您考虑过的其他方案
-      description: 请简明扼要地描述您考虑过的任何其他解决方案或功能
-
-  - type: textarea
-    id: additional
-    attributes:
-      label: 其他补充信息
-      description: 在此添加任何其他与功能建议相关的上下文或截图

.github/ISSUE_TEMPLATE/#2_question.yml

@@ -1,77 +0,0 @@
-name: ❓ 提问 & 讨论 (中文)
-description: 寻求帮助、讨论问题、提出疑问等...
-title: '[讨论]: '
-labels: ['discussion', 'help wanted']
-body:
-  - type: markdown
-    attributes:
-      value: |
-        感谢您的提问！请尽可能详细地描述您的问题，这样我们才能更好地帮助您。
-
-  - type: checkboxes
-    id: checklist
-    attributes:
-      label: Issue 检查清单
-      description: |
-        在提交 Issue 前请确保您已经完成了以下所有步骤
-      options:
-        - label: 我理解 Issue 是用于反馈和解决问题的，而非吐槽评论区，将尽可能提供更多信息帮助问题解决。
-          required: true
-        - label: 我确认自己需要的是提出问题并且讨论问题，而不是 Bug 反馈或需求建议。
-          required: true
-
-  - type: dropdown
-    id: platform
-    attributes:
-      label: 平台
-      description: 您正在使用哪个平台？
-      options:
-        - Windows
-        - macOS
-        - Linux
-    validations:
-      required: true
-
-  - type: input
-    id: version
-    attributes:
-      label: 版本
-      description: 您正在运行的 Cherry Studio 版本是什么？
-      placeholder: 例如 v1.0.0
-    validations:
-      required: true
-
-  - type: textarea
-    id: question
-    attributes:
-      label: 您的问题
-      description: 请详细描述您的问题
-      placeholder: 请尽可能清楚地说明您的问题...
-    validations:
-      required: true
-
-  - type: textarea
-    id: context
-    attributes:
-      label: 相关背景
-      description: 请提供一些背景信息，帮助我们更好地理解您的问题
-      placeholder: 例如：使用场景、已尝试的解决方案等
-
-  - type: textarea
-    id: additional
-    attributes:
-      label: 补充信息
-      description: 任何其他相关的信息、截图或代码示例
-      render: shell
-
-  - type: dropdown
-    id: priority
-    attributes:
-      label: 优先级
-      description: 这个问题对您来说有多紧急？
-      options:
-        - 低 (有空再看)
-        - 中 (希望尽快得到答复)
-        - 高 (阻碍工作进行)
-    validations:
-      required: true

.github/ISSUE_TEMPLATE/#3_others.yml

@@ -1,76 +0,0 @@
-name: 🤔 其他问题 (中文)
-description: 提交不属于错误报告或功能需求的问题
-title: '[其他]: '
-body:
-  - type: markdown
-    attributes:
-      value: |
-        感谢您花时间提出问题！
-        在提交此问题之前，请确保您已经了解了[常见问题](https://docs.cherry-ai.com/question-contact/questions)和[知识科普](https://docs.cherry-ai.com/question-contact/knowledge)
-
-  - type: checkboxes
-    id: checklist
-    attributes:
-      label: 提交前检查
-      description: |
-        在提交 Issue 前请确保您已经完成了以下所有步骤
-      options:
-        - label: 我理解 Issue 是用于反馈和解决问题的，而非吐槽评论区，将尽可能提供更多信息帮助问题解决。
-          required: true
-        - label: 我已经查看了置顶 Issue 并搜索了现有的 [开放Issue](https://github.com/CherryHQ/cherry-studio/issues)和[已关闭Issue](https://github.com/CherryHQ/cherry-studio/issues?q=is%3Aissue%20state%3Aclosed%20)，没有找到类似的问题。
-          required: true
-        - label: 我填写了简短且清晰明确的标题，以便开发者在翻阅 Issue 列表时能快速确定大致问题。而不是"一个问题"、"求助"等。
-          required: true
-        - label: 我的问题不属于错误报告或功能需求类别。
-          required: true
-
-  - type: dropdown
-    id: platform
-    attributes:
-      label: 平台
-      description: 您正在使用哪个平台？
-      options:
-        - Windows
-        - macOS
-        - Linux
-    validations:
-      required: true
-
-  - type: input
-    id: version
-    attributes:
-      label: 版本
-      description: 您正在运行的 Cherry Studio 版本是什么？
-      placeholder: 例如 v1.0.0
-    validations:
-      required: true
-
-  - type: textarea
-    id: question
-    attributes:
-      label: 问题描述
-      description: 请详细描述您的问题或疑问
-      placeholder: 我想了解有关...的更多信息
-    validations:
-      required: true
-
-  - type: textarea
-    id: context
-    attributes:
-      label: 相关背景
-      description: 请提供与您的问题相关的任何背景信息或上下文
-      placeholder: 我尝试实现...时遇到了疑问
-    validations:
-      required: true
-
-  - type: textarea
-    id: attempts
-    attributes:
-      label: 您已尝试的方法
-      description: 请描述您为解决问题已经尝试过的方法（如果有）
-
-  - type: textarea
-    id: additional
-    attributes:
-      label: 附加信息
-      description: 任何能让我们对您的问题有更多了解的信息，包括截图或相关链接

.github/workflows/auto-i18n.yml

@@ -0,0 +1,66 @@
+name: Auto I18N
+
+env:
+  API_KEY: ${{ secrets.TRANSLATE_API_KEY }}
+  MODEL: ${{ vars.MODEL || 'deepseek/deepseek-v3.1'}}
+  BASE_URL: ${{ vars.BASE_URL || 'https://api.ppinfra.com/openai'}}
+
+on:
+  pull_request:
+    types: [opened, synchronize, reopened]
+  workflow_dispatch:
+
+jobs:
+  auto-i18n:
+    runs-on: ubuntu-latest
+    if: github.event.pull_request.head.repo.full_name == 'CherryHQ/cherry-studio'
+    name: Auto I18N
+    permissions:
+      contents: write
+      pull-requests: write
+
+    steps:
+      - name: 🐈‍⬛ Checkout
+        uses: actions/checkout@v5
+        with:
+          ref: ${{ github.event.pull_request.head.ref }}
+
+      - name: 📦 Setting Node.js
+        uses: actions/setup-node@v4
+        with:
+          node-version: 20
+
+      - name: 📦 Install dependencies in isolated directory
+        run: |
+          # 在临时目录安装依赖
+          mkdir -p /tmp/translation-deps
+          cd /tmp/translation-deps
+          echo '{"dependencies": {"openai": "^5.12.2", "cli-progress": "^3.12.0", "tsx": "^4.20.3", "@biomejs/biome": "2.2.4"}}' > package.json
+          npm install --no-package-lock
+
+          # 设置 NODE_PATH 让项目能找到这些依赖
+          echo "NODE_PATH=/tmp/translation-deps/node_modules" >> $GITHUB_ENV
+
+      - name: 🏃‍♀️ Translate
+        run: npx tsx scripts/auto-translate-i18n.ts
+
+      - name: 🔍 Format
+        run: cd /tmp/translation-deps && npx biome format --config-path /home/runner/work/cherry-studio/cherry-studio/biome.jsonc --write /home/runner/work/cherry-studio/cherry-studio/src/renderer/src/i18n/
+
+      - name: 🔄 Commit changes
+        run: |
+          git config --local user.email "action@github.com"
+          git config --local user.name "GitHub Action"
+          git add .
+          git reset -- package.json yarn.lock  # 不提交 package.json 和 yarn.lock 的更改
+          if git diff --cached --quiet; then
+            echo "No changes to commit"
+          else
+            git commit -m "fix(i18n): Auto update translations for PR #${{ github.event.pull_request.number }}"
+          fi
+
+      - name: 🚀 Push changes
+        uses: ad-m/github-push-action@master
+        with:
+          github_token: ${{ secrets.GITHUB_TOKEN }}
+          branch: ${{ github.event.pull_request.head.ref }}

.github/workflows/claude-code-review.yml

@@ -0,0 +1,56 @@
+name: Claude Code Review
+
+on:
+  pull_request:
+    types: [opened]
+    # Optional: Only run on specific file changes
+    # paths:
+    #   - "src/**/*.ts"
+    #   - "src/**/*.tsx"
+    #   - "src/**/*.js"
+    #   - "src/**/*.jsx"
+
+jobs:
+  claude-review:
+    # Only trigger code review for PRs from the main repository due to upstream OIDC issues
+    # https://github.com/anthropics/claude-code-action/issues/542
+    if: |
+      (github.event.pull_request.head.repo.full_name == github.repository) &&
+      (github.event.pull_request.draft == false)
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      pull-requests: write
+      issues: read
+      id-token: write
+      actions: read
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 1
+
+      - name: Run Claude Code Review
+        id: claude-review
+        uses: anthropics/claude-code-action@v1
+        with:
+          claude_code_oauth_token: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
+          prompt: |
+            Please review this pull request and provide feedback on:
+            - Code quality and best practices
+            - Potential bugs or issues
+            - Performance considerations
+            - Security concerns
+            - Test coverage
+
+            PR number: ${{ github.event.number }}
+            Repo: ${{ github.repository }}
+
+            Use the repository's CLAUDE.md for guidance on style and conventions. Be constructive and helpful in your feedback.
+
+            Use `gh pr comment` with your Bash tool to leave your review as a comment on the PR.
+
+          # See https://github.com/anthropics/claude-code-action/blob/main/docs/usage.md
+          # or https://docs.anthropic.com/en/docs/claude-code/sdk#command-line for available options
+          claude_args: '--allowed-tools "Bash(gh issue view:*),Bash(gh search:*),Bash(gh issue list:*),Bash(gh pr comment:*),Bash(gh pr diff:*),Bash(gh pr view:*),Bash(gh pr list:*)"'

.github/workflows/claude-translator.yml

@@ -0,0 +1,107 @@
+name: Claude Translator
+concurrency:
+  group: translator-${{ github.event.comment.id || github.event.issue.number || github.event.review.id }}
+  cancel-in-progress: false
+
+on:
+  issues:
+    types: [opened]
+  issue_comment:
+    types: [created, edited]
+  pull_request_review:
+    types: [submitted, edited]
+  pull_request_review_comment:
+    types: [created, edited]
+
+jobs:
+  translate:
+    if: |
+      (github.event_name == 'issues') ||
+      (github.event_name == 'issue_comment' && github.event.sender.type != 'Bot') ||
+      (github.event_name == 'pull_request_review' && github.event.sender.type != 'Bot') ||
+      (github.event_name == 'pull_request_review_comment' && github.event.sender.type != 'Bot')
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      issues: write # 编辑issues/comments
+      pull-requests: write
+      id-token: write
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 1
+
+      - name: Run Claude for translation
+        uses: anthropics/claude-code-action@main
+        id: claude
+        with:
+          # Warning: Permissions should have been controlled by workflow permission.
+          # Now `contents: read` is safe for files, but we could make a fine-grained token to control it.
+          # See: https://github.com/anthropics/claude-code-action/blob/main/docs/security.md
+          github_token: ${{ secrets.TOKEN_GITHUB_WRITE }}
+          allowed_non_write_users: "*"
+          claude_code_oauth_token: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
+          claude_args: "--allowed-tools Bash(gh issue:*),Bash(gh api:repos/*/issues:*),Bash(gh api:repos/*/pulls/*/reviews/*),Bash(gh api:repos/*/pulls/comments/*)"
+          prompt: |
+            你是一个多语言翻译助手。你需要响应 GitHub Webhooks 中的以下四种事件：
+
+            - issues
+            - issue_comment
+            - pull_request_review
+            - pull_request_review_comment
+
+            请完成以下任务：
+
+            1. 获取当前事件的完整信息。
+
+               - 如果当前事件是 issues，就获取该 issues 的信息。
+               - 如果当前事件是 issue_comment，就获取该 comment 的信息。
+               - 如果当前事件是 pull_request_review，就获取该 review 的信息。
+               - 如果当前事件是 pull_request_review_comment，就获取该 comment 的信息。
+
+            2. 智能检测内容。
+
+                - 如果获取到的信息是已经遵循格式要求翻译过的内容，则检查翻译内容和原始内容是否匹配。若不匹配，则重新翻译一次令其匹配，并遵循格式要求；
+                - 如果获取到的信息是未翻译过的内容，检查其内容语言。若不是英文，则翻译成英文；
+                - 如果获取到的信息是部分翻译为英文的内容，则将其翻译为英文；
+                - 如果获取到的信息包含了对已翻译内容的引用，则将引用内容清理为仅含英文的内容。引用的内容不能够包含"This xxx was translated by Claude"和"Original Content`等内容。
+                - 如果获取到的信息包含了其他类型的引用，即对非 Claude 翻译的内容的引用，则直接照原样引用，不进行翻译。
+                - 如果获取到的信息是通过邮件回复的内容，则在翻译时应当将邮件内容的引用放到最后。在原始内容和翻译内容中只需要回复的内容本身，不要包含对邮件内容的引用。
+                - 如果获取到的信息本身不需要任何处理，则跳过任务。
+
+            3. 格式要求：
+
+               - 标题：英文翻译（如果非英文）
+               - 内容格式：
+                 > [!NOTE]
+                 > This issue/comment/review was translated by Claude.
+
+                 [翻译内容]
+
+                 ---
+                 <details>
+                 <summary>Original Content</summary>
+                 [原始内容]
+                 </details>
+
+            4. 使用gh工具更新：
+
+               - 根据环境信息中的Event类型选择正确的命令：
+                 - 如果 Event 是 'issues': gh issue edit [ISSUE_NUMBER] --title "[英文标题]" --body "[翻译内容 + 原始内容]"
+                 - 如果 Event 是 'issue_comment': gh api -X PATCH /repos/${{ github.repository }}/issues/comments/${{ github.event.comment.id }} -f body="[翻译内容 + 原始内容]"
+                 - 如果 Event 是 'pull_request_review': gh api -X PUT /repos/${{ github.repository }}/pulls/${{ github.event.pull_request.number }}/reviews/${{ github.event.review.id }} -f body="[翻译内容]"
+                 - 如果 Event 是 'pull_request_review_comment': gh api -X PATCH /repos/${{ github.repository }}/pulls/comments/${{ github.event.comment.id }} -f body="[翻译内容 + 原始内容]"
+
+            环境信息：
+            - Event: ${{ github.event_name }}
+            - Issue Number: ${{ github.event.issue.number }}
+            - Repository: ${{ github.repository }}
+            - (Review) Comment ID: ${{ github.event.comment.id || 'N/A' }}
+            - Pull Request Number: ${{ github.event.pull_request.number || 'N/A' }}
+            - Review ID: ${{ github.event.review.id || 'N/A' }}
+
+
+            使用以下命令获取完整信息：
+            gh issue view ${{ github.event.issue.number }} --json title,body,comments

.github/workflows/claude.yml

@@ -0,0 +1,60 @@
+name: Claude Code
+
+on:
+  issue_comment:
+    types: [created]
+  pull_request_review_comment:
+    types: [created]
+  issues:
+    types: [opened]
+  pull_request_review:
+    types: [submitted]
+
+jobs:
+  claude:
+    if: |
+      (github.event_name == 'issue_comment'
+        && contains(github.event.comment.body, '@claude')
+        && contains(fromJSON('["COLLABORATOR","MEMBER","OWNER"]'), github.event.comment.author_association))
+      ||
+      (github.event_name == 'pull_request_review_comment'
+        && contains(github.event.comment.body, '@claude')
+        && contains(fromJSON('["COLLABORATOR","MEMBER","OWNER"]'), github.event.comment.author_association))
+      ||
+      (github.event_name == 'pull_request_review'
+        && contains(github.event.review.body, '@claude')
+        && contains(fromJSON('["COLLABORATOR","MEMBER","OWNER"]'), github.event.review.author_association))
+      ||
+      (github.event_name == 'issues'
+        && (contains(github.event.issue.body, '@claude') || contains(github.event.issue.title, '@claude'))
+        && contains(fromJSON('["COLLABORATOR","MEMBER","OWNER"]'), github.event.issue.author_association))
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      pull-requests: read
+      issues: read
+      id-token: write
+      actions: read # Required for Claude to read CI results on PRs
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 1
+
+      - name: Run Claude Code
+        id: claude
+        uses: anthropics/claude-code-action@v1
+        with:
+          claude_code_oauth_token: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
+
+          # This is an optional setting that allows Claude to read CI results on PRs
+          additional_permissions: |
+            actions: read
+
+          # Optional: Give a custom prompt to Claude. If this is not specified, Claude will perform the instructions specified in the comment that tagged it.
+          # prompt: 'Update the pull request description to include a summary of changes.'
+
+          # Optional: Add claude_args to customize behavior and configuration
+          # See https://github.com/anthropics/claude-code-action/blob/main/docs/usage.md
+          # or https://docs.anthropic.com/en/docs/claude-code/sdk#command-line for available options
+          # claude_args: '--model claude-opus-4-1-20250805 --allowed-tools Bash(gh pr:*)'

.github/workflows/nightly-build.yml

@@ -94,7 +94,6 @@ jobs:
         if: matrix.os == 'ubuntu-latest'
         run: |
           sudo apt-get install -y rpm
-          yarn build:npm linux
           yarn build:linux
         env:
           GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
@@ -107,7 +106,6 @@ jobs:
       - name: Build Mac
         if: matrix.os == 'macos-latest'
         run: |
-          yarn build:npm mac
           yarn build:mac
         env:
           CSC_LINK: ${{ secrets.CSC_LINK }}
@@ -125,7 +123,6 @@ jobs:
       - name: Build Windows
         if: matrix.os == 'windows-latest'
         run: |
-          yarn build:npm windows
           yarn build:win
         env:
           GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/pr-ci.yml

@@ -9,6 +9,7 @@ on:
     branches:
       - main
       - develop
+      - v2
 
 jobs:
   build:
@@ -48,6 +49,9 @@ jobs:
       - name: Lint Check
         run: yarn test:lint
 
+      - name: Format Check
+        run: yarn format:check
+
       - name: Type Check
         run: yarn typecheck
 

.github/workflows/release.yml

@@ -80,7 +80,6 @@ jobs:
         if: matrix.os == 'ubuntu-latest'
         run: |
           sudo apt-get install -y rpm
-          yarn build:npm linux
           yarn build:linux
 
         env:
@@ -95,7 +94,6 @@ jobs:
         if: matrix.os == 'macos-latest'
         run: |
           sudo -H pip install setuptools
-          yarn build:npm mac
           yarn build:mac
         env:
           CSC_LINK: ${{ secrets.CSC_LINK }}
@@ -113,7 +111,6 @@ jobs:
       - name: Build Windows
         if: matrix.os == 'windows-latest'
         run: |
-          yarn build:npm windows
           yarn build:win
         env:
           GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}

.gitignore

@@ -37,6 +37,7 @@ dist
 out
 mcp_server
 stats.html
+.eslintcache
 
 # ENV
 .env
@@ -60,8 +61,13 @@ coverage
 .vitest-cache
 vitest.config.*.timestamp-*
 
+# TypeScript incremental build
+.tsbuildinfo
+
 # playwright
 playwright-report
 test-results
 
 YOUR_MEMORY_FILE_PATH
+
+.sessions/

.oxlintrc.json

@@ -0,0 +1,215 @@
+{
+  "$schema": "./node_modules/oxlint/configuration_schema.json",
+  "categories": {},
+  "env": {
+    "es2022": true
+  },
+  "globals": {},
+  "ignorePatterns": [
+    "node_modules/**",
+    "build/**",
+    "dist/**",
+    "out/**",
+    "local/**",
+    ".yarn/**",
+    ".gitignore",
+    "scripts/cloudflare-worker.js",
+    "src/main/integration/nutstore/sso/lib/**",
+    "src/main/integration/cherryin/index.js",
+    "src/main/integration/nutstore/sso/lib/**",
+    "src/renderer/src/ui/**",
+    "packages/**/dist",
+    "eslint.config.mjs"
+  ],
+  "overrides": [
+    // set different env
+    {
+      "env": {
+        "node": true
+      },
+      "files": ["src/main/**", "resources/scripts/**", "scripts/**", "playwright.config.ts", "electron.vite.config.ts"]
+    },
+    {
+      "env": {
+        "browser": true
+      },
+      "files": [
+        "src/renderer/**/*.{ts,tsx}",
+        "packages/aiCore/**",
+        "packages/extension-table-plus/**",
+        "resources/js/**"
+      ]
+    },
+    {
+      "env": {
+        "node": true,
+        "vitest": true
+      },
+      "files": ["**/__tests__/*.test.{ts,tsx}", "tests/**"]
+    },
+    {
+      "env": {
+        "browser": true,
+        "node": true
+      },
+      "files": ["src/preload/**"]
+    }
+  ],
+  // We don't use the React plugin here because its behavior differs slightly from that of ESLint's React plugin.
+  "plugins": ["unicorn", "typescript", "oxc", "import"],
+  "rules": {
+    "constructor-super": "error",
+    "for-direction": "error",
+    "getter-return": "error",
+    "no-array-constructor": "off",
+    // "import/no-cycle": "error", // tons of error, bro
+    "no-async-promise-executor": "error",
+    "no-caller": "warn",
+    "no-case-declarations": "error",
+    "no-class-assign": "error",
+    "no-compare-neg-zero": "error",
+    "no-cond-assign": "error",
+    "no-const-assign": "error",
+    "no-constant-binary-expression": "error",
+    "no-constant-condition": "error",
+    "no-control-regex": "error",
+    "no-debugger": "error",
+    "no-delete-var": "error",
+    "no-dupe-args": "error",
+    "no-dupe-class-members": "error",
+    "no-dupe-else-if": "error",
+    "no-dupe-keys": "error",
+    "no-duplicate-case": "error",
+    "no-empty": "error",
+    "no-empty-character-class": "error",
+    "no-empty-pattern": "error",
+    "no-empty-static-block": "error",
+    "no-eval": "warn",
+    "no-ex-assign": "error",
+    "no-extra-boolean-cast": "error",
+    "no-fallthrough": "warn",
+    "no-func-assign": "error",
+    "no-global-assign": "error",
+    "no-import-assign": "error",
+    "no-invalid-regexp": "error",
+    "no-irregular-whitespace": "error",
+    "no-loss-of-precision": "error",
+    "no-misleading-character-class": "error",
+    "no-new-native-nonconstructor": "error",
+    "no-nonoctal-decimal-escape": "error",
+    "no-obj-calls": "error",
+    "no-octal": "error",
+    "no-prototype-builtins": "error",
+    "no-redeclare": "error",
+    "no-regex-spaces": "error",
+    "no-self-assign": "error",
+    "no-setter-return": "error",
+    "no-shadow-restricted-names": "error",
+    "no-sparse-arrays": "error",
+    "no-this-before-super": "error",
+    "no-unassigned-vars": "warn",
+    "no-undef": "error",
+    "no-unexpected-multiline": "error",
+    "no-unreachable": "error",
+    "no-unsafe-finally": "error",
+    "no-unsafe-negation": "error",
+    "no-unsafe-optional-chaining": "error",
+    "no-unused-expressions": "off", // this rule disallow us to use expression to call function, like `condition && fn()`
+    "no-unused-labels": "error",
+    "no-unused-private-class-members": "error",
+    "no-unused-vars": ["error", { "caughtErrors": "none" }],
+    "no-useless-backreference": "error",
+    "no-useless-catch": "error",
+    "no-useless-escape": "error",
+    "no-useless-rename": "warn",
+    "no-with": "error",
+    "oxc/bad-array-method-on-arguments": "warn",
+    "oxc/bad-char-at-comparison": "warn",
+    "oxc/bad-comparison-sequence": "warn",
+    "oxc/bad-min-max-func": "warn",
+    "oxc/bad-object-literal-comparison": "warn",
+    "oxc/bad-replace-all-arg": "warn",
+    "oxc/const-comparisons": "warn",
+    "oxc/double-comparisons": "warn",
+    "oxc/erasing-op": "warn",
+    "oxc/missing-throw": "warn",
+    "oxc/number-arg-out-of-range": "warn",
+    "oxc/only-used-in-recursion": "off", // manually off bacause of existing warning. may turn it on in the future
+    "oxc/uninvoked-array-callback": "warn",
+    "require-yield": "error",
+    "typescript/await-thenable": "warn",
+    // "typescript/ban-ts-comment": "error",
+    "typescript/no-array-constructor": "error",
+    // "typescript/consistent-type-imports": "error",
+    "typescript/no-array-delete": "warn",
+    "typescript/no-base-to-string": "warn",
+    "typescript/no-duplicate-enum-values": "error",
+    "typescript/no-duplicate-type-constituents": "warn",
+    "typescript/no-empty-object-type": "off",
+    "typescript/no-explicit-any": "off", // not safe but too many errors
+    "typescript/no-extra-non-null-assertion": "error",
+    "typescript/no-floating-promises": "warn",
+    "typescript/no-for-in-array": "warn",
+    "typescript/no-implied-eval": "warn",
+    "typescript/no-meaningless-void-operator": "warn",
+    "typescript/no-misused-new": "error",
+    "typescript/no-misused-spread": "warn",
+    "typescript/no-namespace": "error",
+    "typescript/no-non-null-asserted-optional-chain": "off", // it's off now. but may turn it on.
+    "typescript/no-redundant-type-constituents": "warn",
+    "typescript/no-require-imports": "off",
+    "typescript/no-this-alias": "error",
+    "typescript/no-unnecessary-parameter-property-assignment": "warn",
+    "typescript/no-unnecessary-type-constraint": "error",
+    "typescript/no-unsafe-declaration-merging": "error",
+    "typescript/no-unsafe-function-type": "error",
+    "typescript/no-unsafe-unary-minus": "warn",
+    "typescript/no-useless-empty-export": "warn",
+    "typescript/no-wrapper-object-types": "error",
+    "typescript/prefer-as-const": "error",
+    "typescript/prefer-namespace-keyword": "error",
+    "typescript/require-array-sort-compare": "warn",
+    "typescript/restrict-template-expressions": "warn",
+    "typescript/triple-slash-reference": "error",
+    "typescript/unbound-method": "warn",
+    "unicorn/no-await-in-promise-methods": "warn",
+    "unicorn/no-empty-file": "off", // manually off bacause of existing warning. may turn it on in the future
+    "unicorn/no-invalid-fetch-options": "warn",
+    "unicorn/no-invalid-remove-event-listener": "warn",
+    "unicorn/no-new-array": "off", // manually off bacause of existing warning. may turn it on in the future
+    "unicorn/no-single-promise-in-promise-methods": "warn",
+    "unicorn/no-thenable": "off", // manually off bacause of existing warning. may turn it on in the future
+    "unicorn/no-unnecessary-await": "warn",
+    "unicorn/no-useless-fallback-in-spread": "warn",
+    "unicorn/no-useless-length-check": "warn",
+    "unicorn/no-useless-spread": "off", // manually off bacause of existing warning. may turn it on in the future
+    "unicorn/prefer-set-size": "warn",
+    "unicorn/prefer-string-starts-ends-with": "warn",
+    "use-isnan": "error",
+    "valid-typeof": "error"
+  },
+  "settings": {
+    "jsdoc": {
+      "augmentsExtendsReplacesDocs": false,
+      "exemptDestructuredRootsFromChecks": false,
+      "ignoreInternal": false,
+      "ignorePrivate": false,
+      "ignoreReplacesDocs": true,
+      "implementsReplacesDocs": false,
+      "overrideReplacesDocs": true,
+      "tagNamePreference": {}
+    },
+    "jsx-a11y": {
+      "attributes": {},
+      "components": {},
+      "polymorphicPropName": null
+    },
+    "next": {
+      "rootDir": []
+    },
+    "react": {
+      "formComponents": [],
+      "linkComponents": []
+    }
+  }
+}

.prettierignore

@@ -1,10 +0,0 @@
-out
-dist
-pnpm-lock.yaml
-LICENSE.md
-tsconfig.json
-tsconfig.*.json
-CHANGELOG*.md
-agents.json
-src/renderer/src/integration/nutstore/sso/lib
-src/main/integration/cherryin/index.js

.prettierrc

@@ -1,11 +0,0 @@
-{
-  "bracketSameLine": true,
-  "endOfLine": "lf",
-  "jsonRecursiveSort": true,
-  "jsonSortOrder": "{\"*\": \"lexical\"}",
-  "plugins": ["prettier-plugin-sort-json"],
-  "printWidth": 120,
-  "semi": false,
-  "singleQuote": true,
-  "trailingComma": "none"
-}

.vscode/extensions.json

@@ -1,8 +1,11 @@
 {
   "recommendations": [
     "dbaeumer.vscode-eslint",
-    "esbenp.prettier-vscode",
     "editorconfig.editorconfig",
-    "lokalise.i18n-ally"
+    "lokalise.i18n-ally",
+    "bradlc.vscode-tailwindcss",
+    "vitest.explorer",
+    "oxc.oxc-vscode",
+    "biomejs.biome"
   ]
 }

.vscode/settings.json

@@ -1,38 +1,43 @@
 {
   "[css]": {
-    "editor.defaultFormatter": "esbenp.prettier-vscode"
+    "editor.defaultFormatter": "biomejs.biome"
   },
   "[javascript]": {
-    "editor.defaultFormatter": "esbenp.prettier-vscode"
+    "editor.defaultFormatter": "biomejs.biome"
   },
   "[json]": {
-    "editor.defaultFormatter": "esbenp.prettier-vscode"
+    "editor.defaultFormatter": "biomejs.biome"
   },
   "[jsonc]": {
-    "editor.defaultFormatter": "esbenp.prettier-vscode"
+    "editor.defaultFormatter": "biomejs.biome"
   },
   "[markdown]": {
     "files.trimTrailingWhitespace": false
   },
   "[scss]": {
-    "editor.defaultFormatter": "esbenp.prettier-vscode"
+    "editor.defaultFormatter": "biomejs.biome"
   },
   "[typescript]": {
-    "editor.defaultFormatter": "esbenp.prettier-vscode"
+    "editor.defaultFormatter": "biomejs.biome"
   },
   "[typescriptreact]": {
-    "editor.defaultFormatter": "esbenp.prettier-vscode"
+    "editor.defaultFormatter": "biomejs.biome"
   },
   "editor.codeActionsOnSave": {
+    "source.fixAll.biome": "explicit",
     "source.fixAll.eslint": "explicit",
+    "source.fixAll.oxc": "explicit",
     "source.organizeImports": "never"
   },
   "editor.formatOnSave": true,
+  "files.associations": {
+    "*.css": "tailwindcss"
+  },
   "files.eol": "\n",
-  "i18n-ally.displayLanguage": "zh-cn",
+  // "i18n-ally.displayLanguage": "zh-cn", // 界面显示语言
   "i18n-ally.enabledFrameworks": ["react-i18next", "i18next"],
   "i18n-ally.enabledParsers": ["ts", "js", "json"], // 解析语言
-  "i18n-ally.fullReloadOnChanged": true, // 界面显示语言
+  "i18n-ally.fullReloadOnChanged": true,
   "i18n-ally.keystyle": "nested", // 翻译路径格式
   "i18n-ally.localesPaths": ["src/renderer/src/i18n/locales"],
   // "i18n-ally.namespace": true, // 开启命名空间

.yarn/patches/@anthropic-ai-claude-code-npm-1.0.118-bbf4e9e59f.patch

@@ -0,0 +1,18 @@
+diff --git a/sdk.mjs b/sdk.mjs
+index e2dbafb4e2faa1bf2b6b02f0009a2b9bbf57c757..3f07a1d5c2949a246fe5414e69ab45942fa605a2 100644
+--- a/sdk.mjs
++++ b/sdk.mjs
+@@ -6355,11 +6355,11 @@ class ProcessTransport {
+         prompt,
+         additionalDirectories = [],
+         cwd,
+-        executable = isRunningWithBun() ? "bun" : "node",
++        executable = process.execPath,
+         executableArgs = [],
+         extraArgs = {},
+         pathToClaudeCodeExecutable,
+-        env = { ...process.env },
++        env = { ...process.env, ELECTRON_RUN_AS_NODE: '1' },
+         stderr,
+         customSystemPrompt,
+         appendSystemPrompt,

.yarn/patches/@napi-rs-system-ocr-npm-1.0.2-59e7a78e8b.patch

@@ -0,0 +1,30 @@
+diff --git a/index.js b/index.js
+index dc071739e79876dff88e1be06a9168e294222d13..b9df7525c62bdf777e89e732e1b0c81f84d872f2 100644
+--- a/index.js
++++ b/index.js
+@@ -380,7 +380,7 @@ if (!nativeBinding || process.env.NAPI_RS_FORCE_WASI) {
+   }
+ }
+ 
+-if (!nativeBinding) {
++if (!nativeBinding && process.platform !== 'linux') {
+   if (loadErrors.length > 0) {
+     throw new Error(
+       `Cannot find native binding. ` +
+@@ -392,6 +392,13 @@ if (!nativeBinding) {
+   throw new Error(`Failed to load native binding`)
+ }
+ 
+-module.exports = nativeBinding
+-module.exports.OcrAccuracy = nativeBinding.OcrAccuracy
+-module.exports.recognize = nativeBinding.recognize
++if (process.platform === 'linux') {
++  module.exports = {OcrAccuracy: {
++    Fast: 0,
++    Accurate: 1
++  }, recognize: () => Promise.resolve({text: '', confidence: 1.0})}
++}else{
++  module.exports = nativeBinding
++  module.exports.OcrAccuracy = nativeBinding.OcrAccuracy
++  module.exports.recognize = nativeBinding.recognize
++}

.yarnrc.yml

@@ -5,3 +5,5 @@ httpTimeout: 300000
 nodeLinker: node-modules
 
 yarnPath: .yarn/releases/yarn-4.9.1.cjs
+npmRegistryServer: https://registry.npmjs.org
+npmPublishRegistry: https://registry.npmjs.org

AGENTS.md

@@ -0,0 +1 @@
+CLAUDE.md

CLAUDE.md

@@ -1,120 +1,52 @@
-# CLAUDE.md
+# AI Assistant Guide
 
-This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
+This file provides guidance to AI coding assistants when working with code in this repository. Adherence to these guidelines is crucial for maintaining code quality and consistency.
+
+## Guiding Principles
+
+- **Clarity and Simplicity**: Write code that is easy to understand and maintain.
+- **Consistency**: Follow existing patterns and conventions in the codebase.
+- **Correctness**: Ensure code is correct, well-tested, and robust.
+- **Efficiency**: Write performant code and use resources judiciously.
+
+## MUST Follow Rules
+
+1.  **Code Search**: Use `ast-grep` for semantic code pattern searches when available. Fallback to `rg` (ripgrep) or `grep` for text-based searches.
+2.  **UI Framework**: Exclusively use **HeroUI** for all new UI components. The use of `antd` or `styled-components` is strictly **PROHIBITED**.
+3.  **Quality Assurance**: **Always** run `yarn build:check` before finalizing your work or making any commits. This ensures code quality (linting, testing, and type checking).
+4.  **Centralized Logging**: Use the `loggerService` exclusively for all application logging (info, warn, error levels) with proper context. Do not use `console.log`.
+5.  **External Research**: Leverage `subagent` for gathering external information, including latest documentation, API references, news, or web-based research. This keeps the main conversation focused on the task at hand.
+6.  **Code Reviews**: Always seek a code review from a human developer before merging significant changes. This ensures adherence to project standards and catches potential issues.
+7.  **Documentation**: Update or create documentation for any new features, modules, or significant changes to existing functionality.
 
 ## Development Commands
 
-### Environment Setup
+- **Install**: `yarn install`
+- **Development**: `yarn dev` - Runs Electron app in development mode
+- **Debug**: `yarn debug` - Starts with debugging enabled, use chrome://inspect
+- **Build Check**: `yarn build:check` - REQUIRED before commits (lint + test + typecheck)
+- **Test**: `yarn test` - Run all tests (Vitest)
+- **Single Test**: `yarn test:main` or `yarn test:renderer`
+- **Lint**: `yarn lint` - Fix linting issues and run typecheck
 
-- **Prerequisites**: Node.js v22.x.x or higher, Yarn 4.9.1
+## Project Architecture
-- **Setup Yarn**: `corepack enable && corepack prepare yarn@4.9.1 --activate`
-- **Install Dependencies**: `yarn install`
 
-### Development
+### Electron Structure
+- **Main Process** (`src/main/`): Node.js backend with services (MCP, Knowledge, Storage, etc.)
+- **Renderer Process** (`src/renderer/`): React UI with Redux state management
+- **Preload Scripts** (`src/preload/`): Secure IPC bridge
 
-- **Start Development**: `yarn dev` - Runs Electron app in development mode
+### Key Components
-- **Debug Mode**: `yarn debug` - Starts with debugging enabled, use chrome://inspect
+- **AI Core** (`src/renderer/src/aiCore/`): Middleware pipeline for multiple AI providers.
-
+- **Services** (`src/main/services/`): MCPService, KnowledgeService, WindowService, etc.
-### Testing & Quality
+- **Build System**: Electron-Vite with experimental rolldown-vite, yarn workspaces.
-
+- **State Management**: Redux Toolkit (`src/renderer/src/store/`) for predictable state.
-- **Run Tests**: `yarn test` - Runs all tests (Vitest)
+- **UI Components**: HeroUI (`@heroui/*`) for all new UI elements.
-- **Run E2E Tests**: `yarn test:e2e` - Playwright end-to-end tests
-- **Type Check**: `yarn typecheck` - Checks TypeScript for both node and web
-- **Lint**: `yarn lint` - ESLint with auto-fix
-- **Format**: `yarn format` - Prettier formatting
-
-### Build & Release
-
-- **Build**: `yarn build` - Builds for production (includes typecheck)
-- **Platform-specific builds**:
-  - Windows: `yarn build:win`
-  - macOS: `yarn build:mac`
-  - Linux: `yarn build:linux`
-
-## Architecture Overview
-
-### Electron Multi-Process Architecture
-
-- **Main Process** (`src/main/`): Node.js backend handling system integration, file operations, and services
-- **Renderer Process** (`src/renderer/`): React-based UI running in Chromium
-- **Preload Scripts** (`src/preload/`): Secure bridge between main and renderer processes
-
-### Key Architectural Components
-
-#### Main Process Services (`src/main/services/`)
-
-- **MCPService**: Model Context Protocol server management
-- **KnowledgeService**: Document processing and knowledge base management
-- **FileStorage/S3Storage/WebDav**: Multiple storage backends
-- **WindowService**: Multi-window management (main, mini, selection windows)
-- **ProxyManager**: Network proxy handling
-- **SearchService**: Full-text search capabilities
-
-#### AI Core (`src/renderer/src/aiCore/`)
-
-- **Middleware System**: Composable pipeline for AI request processing
-- **Client Factory**: Supports multiple AI providers (OpenAI, Anthropic, Gemini, etc.)
-- **Stream Processing**: Real-time response handling
-
-#### State Management (`src/renderer/src/store/`)
-
-- **Redux Toolkit**: Centralized state management
-- **Persistent Storage**: Redux-persist for data persistence
-- **Thunks**: Async actions for complex operations
-
-#### Knowledge Management
-
-- **Embeddings**: Vector search with multiple providers (OpenAI, Voyage, etc.)
-- **OCR**: Document text extraction (system OCR, Doc2x, Mineru)
-- **Preprocessing**: Document preparation pipeline
-- **Loaders**: Support for various file formats (PDF, DOCX, EPUB, etc.)
-
-### Build System
-
-- **Electron-Vite**: Development and build tooling (v4.0.0)
-- **Rolldown-Vite**: Using experimental rolldown-vite instead of standard vite
-- **Workspaces**: Monorepo structure with `packages/` directory
-- **Multiple Entry Points**: Main app, mini window, selection toolbar
-- **Styled Components**: CSS-in-JS styling with SWC optimization
-
-### Testing Strategy
-
-- **Vitest**: Unit and integration testing
-- **Playwright**: End-to-end testing
-- **Component Testing**: React Testing Library
-- **Coverage**: Available via `yarn test:coverage`
-
-### Key Patterns
-
-- **IPC Communication**: Secure main-renderer communication via preload scripts
-- **Service Layer**: Clear separation between UI and business logic
-- **Plugin Architecture**: Extensible via MCP servers and middleware
-- **Multi-language Support**: i18n with dynamic loading
-- **Theme System**: Light/dark themes with custom CSS variables
-
-## Logging Standards
-
-### Usage
 
+### Logging
 ```typescript
-// Main process
 import { loggerService } from '@logger'
 const logger = loggerService.withContext('moduleName')
-
+// Renderer: loggerService.initWindowSource('windowName') first
-// Renderer process (set window source first)
-loggerService.initWindowSource('windowName')
-const logger = loggerService.withContext('moduleName')
-
-// Logging
 logger.info('message', CONTEXT)
-logger.error('message', new Error('error'), CONTEXT)
 ```
-
-### Log Levels (highest to lowest)
-
-- `error` - Critical errors causing crash/unusable functionality
-- `warn` - Potential issues that don't affect core functionality
-- `info` - Application lifecycle and key user actions
-- `verbose` - Detailed flow information for feature tracing
-- `debug` - Development diagnostic info (not for production)
-- `silly` - Extreme debugging, low-level information

LICENSE

@@ -1,48 +1,3 @@
-**许可协议 (Licensing)**
-
-本项目采用**区分用户的双重许可 (User-Segmented Dual Licensing)** 模式。
-
-**核心原则：**
-
-* **个人用户 和 10人及以下企业/组织:** 默认适用 **GNU Affero 通用公共许可证 v3.0 (AGPLv3)**。  
-* **超过10人的企业/组织:** **必须** 获取 **商业许可证 (Commercial License)**。
-
-定义：“10人及以下”  
-指在您的组织（包括公司、非营利组织、政府机构、教育机构等任何实体）中，能够访问、使用或以任何方式直接或间接受益于本软件（Cherry Studio）功能的个人总数不超过10人。这包括但不限于开发者、测试人员、运营人员、最终用户、通过集成系统间接使用者等。
-
----
-
-**1. 开源许可证 (Open Source License): AGPLv3 - 适用于个人及10人及以下组织**
-
-* 如果您是个人用户，或者您的组织满足上述“10人及以下”的定义，您可以在 **AGPLv3** 的条款下自由使用、修改和分发 Cherry Studio。AGPLv3 的完整文本可以访问 [https://www.gnu.org/licenses/agpl-3.0.html](https://www.gnu.org/licenses/agpl-3.0.html) 获取。  
-* **核心义务：** AGPLv3 的一个关键要求是，如果您修改了 Cherry Studio 并通过网络提供服务，或者分发了修改后的版本，您必须以 AGPLv3 许可证向接收者提供相应的**完整源代码**。即使您符合“10人及以下”的标准，如果您希望避免此源代码公开义务，您也需要考虑获取商业许可证（见下文）。  
-* 使用前请务必仔细阅读并理解 AGPLv3 的所有条款。
-
-**2. 商业许可证 (Commercial License) - 适用于超过10人的组织，或希望规避 AGPLv3 义务的用户**
-
-* **强制要求：** 如果您的组织**不**满足上述“10人及以下”的定义（即有11人或更多人可以访问、使用或受益于本软件），您**必须**联系我们获取并签署一份商业许可证才能使用 Cherry Studio。  
-* **自愿选择：** 即使您的组织满足“10人及以下”的条件，但如果您的使用场景**无法满足 AGPLv3 的条款要求**（特别是关于**源代码公开**的义务），或者您需要 AGPLv3 **未提供**的特定商业条款（如保证、赔偿、无 Copyleft 限制等），您也**必须**联系我们获取并签署一份商业许可证。  
-* **需要商业许可证的常见情况包括（但不限于）：**  
-  * 您的组织规模超过10人。  
-  * （无论组织规模）您希望分发修改过的 Cherry Studio 版本，但**不希望**根据 AGPLv3 公开您修改部分的源代码。  
-  * （无论组织规模）您希望基于修改过的 Cherry Studio 提供网络服务（SaaS），但**不希望**根据 AGPLv3 向服务使用者提供修改后的源代码。  
-  * （无论组织规模）您的公司政策、客户合同或项目要求不允许使用 AGPLv3 许可的软件，或要求闭源分发及保密。  
-* 商业许可证将为您提供豁免 AGPLv3 义务（如源代码公开）的权利，并可能包含额外的商业保障条款。  
-* **获取商业许可：** 请通过邮箱 **bd@cherry-ai.com** 联系 Cherry Studio 开发团队洽谈商业授权事宜。
-
-**3. 贡献 (Contributions)**
-
-* 我们欢迎社区对 Cherry Studio 的贡献。所有向本项目提交的贡献都将被视为在 **AGPLv3** 许可证下提供。  
-* 通过向本项目提交贡献（例如通过 Pull Request），即表示您同意您的代码以 AGPLv3 许可证授权给本项目及所有后续使用者（无论这些使用者最终遵循 AGPLv3 还是商业许可）。  
-* 您也理解并同意，您的贡献可能会被包含在根据商业许可证分发的 Cherry Studio 版本中。
-
-**4. 其他条款 (Other Terms)**
-
-* 关于商业许可证的具体条款和条件，以双方签署的正式商业许可协议为准。  
-* 项目维护者保留根据需要更新本许可政策（包括用户规模定义和阈值）的权利。相关更新将通过项目官方渠道（如代码仓库、官方网站）进行通知。
-
----
-
 **Licensing**
 
 This project employs a **User-Segmented Dual Licensing** model.

README.md

@@ -82,7 +82,7 @@ Cherry Studio is a desktop client that supports multiple LLM providers, availabl
 1. **Diverse LLM Provider Support**:
 
 - ☁️ Major LLM Cloud Services: OpenAI, Gemini, Anthropic, and more
-- 🔗 AI Web Service Integration: Claude, Peplexity, Poe, and others
+- 🔗 AI Web Service Integration: Claude, Perplexity, Poe, and others
 - 💻 Local Model Support with Ollama, LM Studio
 
 2. **AI Assistants & Conversations**:

biome.jsonc

@@ -0,0 +1,97 @@
+{
+  "$schema": "https://biomejs.dev/schemas/2.2.4/schema.json",
+  "assist": {
+    // to sort json
+    "actions": {
+      "source": {
+        "organizeImports": "on",
+        "useSortedKeys": {
+          "level": "on",
+          "options": {
+            "sortOrder": "lexicographic"
+          }
+        }
+      }
+    },
+    "enabled": true,
+    "includes": ["**/*.json", "!*.json", "!**/package.json"]
+  },
+  "css": {
+    "formatter": {
+      "quoteStyle": "single"
+    }
+  },
+  "files": { "ignoreUnknown": false },
+  "formatter": {
+    "attributePosition": "auto",
+    "bracketSameLine": false,
+    "bracketSpacing": true,
+    "enabled": true,
+    "expand": "auto",
+    "formatWithErrors": true,
+    "includes": [
+      "**",
+      "!out/**",
+      "!**/dist/**",
+      "!build/**",
+      "!.yarn/**",
+      "!.github/**",
+      "!.husky/**",
+      "!.vscode/**",
+      "!*.yaml",
+      "!*.yml",
+      "!*.mjs",
+      "!*.cjs",
+      "!*.md",
+      "!*.json",
+      "!src/main/integration/**",
+      "!**/tailwind.css",
+      "!**/package.json"
+    ],
+    "indentStyle": "space",
+    "indentWidth": 2,
+    "lineEnding": "lf",
+    "lineWidth": 120,
+    "useEditorconfig": true
+  },
+  "html": { "formatter": { "selfCloseVoidElements": "always" } },
+  "javascript": {
+    "formatter": {
+      "arrowParentheses": "always",
+      "attributePosition": "auto",
+      // To minimize changes in this PR as much as possible, it's set to true. However, setting it to false would make it more convenient to add attributes at the end.
+      "bracketSameLine": true,
+      "bracketSpacing": true,
+      "jsxQuoteStyle": "double",
+      "quoteProperties": "asNeeded",
+      "quoteStyle": "single",
+      "semicolons": "asNeeded",
+      "trailingCommas": "none"
+    }
+  },
+  "json": {
+    "parser": {
+      "allowComments": true
+    }
+  },
+  "linter": {
+    "enabled": true,
+    "includes": ["!**/tailwind.css", "src/renderer/**/*.{tsx,ts}"],
+    // only enable sorted tailwind css rule. used as formatter instead of linter
+    "rules": {
+      "nursery": {
+        // to sort tailwind css classes
+        "useSortedClasses": {
+          "fix": "safe",
+          "level": "warn",
+          "options": {
+            "functions": ["cn"]
+          }
+        }
+      },
+      "recommended": false,
+      "suspicious": "off"
+    }
+  },
+  "vcs": { "clientKind": "git", "enabled": false, "useIgnoreFile": false }
+}

build/entitlements.mac.plist

@@ -8,5 +8,7 @@
     <true/>
     <key>com.apple.security.cs.allow-dyld-environment-variables</key>
     <true/>
+    <key>com.apple.security.cs.disable-library-validation</key>
+    <true/>
   </dict>
 </plist>

components.json

@@ -0,0 +1,21 @@
+{
+  "$schema": "https://ui.shadcn.com/schema.json",
+  "aliases": {
+    "components": "@renderer/ui/third-party",
+    "hooks": "@renderer/hooks",
+    "lib": "@renderer/lib",
+    "ui": "@renderer/ui",
+    "utils": "@renderer/utils"
+  },
+  "iconLibrary": "lucide",
+  "rsc": false,
+  "style": "new-york",
+  "tailwind": {
+    "baseColor": "zinc",
+    "config": "",
+    "css": "src/renderer/src/assets/styles/tailwind.css",
+    "cssVariables": true,
+    "prefix": ""
+  },
+  "tsx": true
+}

docs/README.zh.md

@@ -13,7 +13,7 @@
         <p><a href="https://openaitx.github.io/view.html?user=CherryHQ&project=cherry-studio&lang=fr">Français</a></p>
         <p><a href="https://openaitx.github.io/view.html?user=CherryHQ&project=cherry-studio&lang=de">Deutsch</a></p>
         <p><a href="https://openaitx.github.io/view.html?user=CherryHQ&project=cherry-studio&lang=es">Español</a></p>
-        <p><a href="https://openaitx.github.io/view.html?user=CherryHQ&project=cherry-studio&lang=it">Itapano</a></p>
+        <p><a href="https://openaitx.github.io/view.html?user=CherryHQ&project=cherry-studio&lang=it">Italiano</a></p>
         <p><a href="https://openaitx.github.io/view.html?user=CherryHQ&project=cherry-studio&lang=ru">Русский</a></p>
         <p><a href="https://openaitx.github.io/view.html?user=CherryHQ&project=cherry-studio&lang=pt">Português</a></p>
         <p><a href="https://openaitx.github.io/view.html?user=CherryHQ&project=cherry-studio&lang=nl">Nederlands</a></p>
@@ -89,7 +89,7 @@ https://docs.cherry-ai.com
 1. **多样化 LLM 服务支持**：
 
 - ☁️ 支持主流 LLM 云服务：OpenAI、Gemini、Anthropic、硅基流动等
-- 🔗 集成流行 AI Web 服务：Claude、Peplexity、Poe、腾讯元宝、知乎直答等
+- 🔗 集成流行 AI Web 服务：Claude、Perplexity、Poe、腾讯元宝、知乎直答等
 - 💻 支持 Ollama、LM Studio 本地模型部署
 
 2. **智能助手与对话**：

docs/dev.md

@@ -2,7 +2,9 @@
 
 ## IDE Setup
 
-[Cursor](https://www.cursor.com/) + [ESLint](https://marketplace.visualstudio.com/items?itemName=dbaeumer.vscode-eslint) + [Prettier](https://marketplace.visualstudio.com/items?itemName=esbenp.prettier-vscode)
+- Editor: [Cursor](https://www.cursor.com/), etc. Any VS Code compatible editor.
+- Linter: [ESLint](https://marketplace.visualstudio.com/items?itemName=dbaeumer.vscode-eslint)
+- Formatter: [Biome](https://marketplace.visualstudio.com/items?itemName=biomejs.biome)
 
 ## Project Setup
 

docs/technical/db.translate_languages.md

@@ -8,9 +8,9 @@
 
 | 字段名     | 类型   | 是否主键 | 索引 | 说明                                                                     |
 | ---------- | ------ | -------- | ---- | ------------------------------------------------------------------------ |
-| `id`       | string | ✅ 是     | ✅    | 唯一标识符，主键                                                         |
+| `id`       | string | ✅ 是    | ✅   | 唯一标识符，主键                                                         |
-| `langCode` | string | ❌ 否     | ✅    | 语言代码（如：`zh-cn`, `en-us`, `ja-jp` 等，均为小写），支持普通索引查询 |
+| `langCode` | string | ❌ 否    | ✅   | 语言代码（如：`zh-cn`, `en-us`, `ja-jp` 等，均为小写），支持普通索引查询 |
-| `value`    | string | ❌ 否     | ❌    | 语言的名称，用户输入                                                     |
+| `value`    | string | ❌ 否    | ❌   | 语言的名称，用户输入                                                     |
-| `emoji`    | string | ❌ 否     | ❌    | 语言的emoji，用户输入                                                    |
+| `emoji`    | string | ❌ 否    | ❌   | 语言的emoji，用户输入                                                    |
 
 > `langCode` 虽非主键，但在业务层应当避免重复插入相同语言代码。

electron-builder.yml

@@ -17,51 +17,52 @@ protocols:
     schemes:
       - cherrystudio
 files:
-  - '**/*'
+  - "**/*"
-  - '!**/{.vscode,.yarn,.yarn-lock,.github,.cursorrules,.prettierrc}'
+  - "!**/{.vscode,.yarn,.yarn-lock,.github,.cursorrules,.prettierrc}"
-  - '!electron.vite.config.{js,ts,mjs,cjs}}'
+  - "!electron.vite.config.{js,ts,mjs,cjs}}"
-  - '!**/{.eslintignore,.eslintrc.js,.eslintrc.json,.eslintcache,root.eslint.config.js,eslint.config.js,.eslintrc.cjs,.prettierignore,.prettierrc.yaml,eslint.config.mjs,dev-app-update.yml,CHANGELOG.md,README.md}'
+  - "!**/{.eslintignore,.eslintrc.js,.eslintrc.json,.eslintcache,root.eslint.config.js,eslint.config.js,.eslintrc.cjs,.prettierignore,.prettierrc.yaml,eslint.config.mjs,dev-app-update.yml,CHANGELOG.md,README.md,biome.jsonc}"
-  - '!**/{.env,.env.*,.npmrc,pnpm-lock.yaml}'
+  - "!**/{.env,.env.*,.npmrc,pnpm-lock.yaml}"
-  - '!**/{tsconfig.json,tsconfig.tsbuildinfo,tsconfig.node.json,tsconfig.web.json}'
+  - "!**/{tsconfig.json,tsconfig.tsbuildinfo,tsconfig.node.json,tsconfig.web.json}"
-  - '!**/{.editorconfig,.jekyll-metadata}'
+  - "!**/{.editorconfig,.jekyll-metadata}"
-  - '!src'
+  - "!src"
-  - '!scripts'
+  - "!scripts"
-  - '!local'
+  - "!local"
-  - '!docs'
+  - "!docs"
-  - '!packages'
+  - "!packages"
-  - '!.swc'
+  - "!.swc"
-  - '!.bin'
+  - "!.bin"
-  - '!._*'
+  - "!._*"
-  - '!*.log'
+  - "!*.log"
-  - '!stats.html'
+  - "!stats.html"
-  - '!*.md'
+  - "!*.md"
-  - '!**/*.{iml,o,hprof,orig,pyc,pyo,rbc,swp,csproj,sln,xproj}'
+  - "!**/*.{iml,o,hprof,orig,pyc,pyo,rbc,swp,csproj,sln,xproj}"
-  - '!**/*.{map,ts,tsx,jsx,less,scss,sass,css.d.ts,d.cts,d.mts,md,markdown,yaml,yml}'
+  - "!**/*.{map,ts,tsx,jsx,less,scss,sass,css.d.ts,d.cts,d.mts,md,markdown,yaml,yml}"
-  - '!**/{test,tests,__tests__,powered-test,coverage}/**'
+  - "!**/{test,tests,__tests__,powered-test,coverage}/**"
-  - '!**/{example,examples}/**'
+  - "!**/{example,examples}/**"
-  - '!**/*.{spec,test}.{js,jsx,ts,tsx}'
+  - "!**/*.{spec,test}.{js,jsx,ts,tsx}"
-  - '!**/*.min.*.map'
+  - "!**/*.min.*.map"
-  - '!**/*.d.ts'
+  - "!**/*.d.ts"
-  - '!**/dist/es6/**'
+  - "!**/dist/es6/**"
-  - '!**/dist/demo/**'
+  - "!**/dist/demo/**"
-  - '!**/amd/**'
+  - "!**/amd/**"
-  - '!**/{.DS_Store,Thumbs.db,thumbs.db,__pycache__}'
+  - "!**/{.DS_Store,Thumbs.db,thumbs.db,__pycache__}"
-  - '!**/{LICENSE,license,LICENSE.*,*.LICENSE.txt,NOTICE.txt,README.md,readme.md,CHANGELOG.md}'
+  - "!**/{LICENSE,license,LICENSE.*,*.LICENSE.txt,NOTICE.txt,README.md,readme.md,CHANGELOG.md}"
-  - '!node_modules/rollup-plugin-visualizer'
+  - "!node_modules/rollup-plugin-visualizer"
-  - '!node_modules/js-tiktoken'
+  - "!node_modules/js-tiktoken"
-  - '!node_modules/@tavily/core/node_modules/js-tiktoken'
+  - "!node_modules/@tavily/core/node_modules/js-tiktoken"
-  - '!node_modules/pdf-parse/lib/pdf.js/{v1.9.426,v1.10.88,v2.0.550}'
+  - "!node_modules/pdf-parse/lib/pdf.js/{v1.9.426,v1.10.88,v2.0.550}"
-  - '!node_modules/mammoth/{mammoth.browser.js,mammoth.browser.min.js}'
+  - "!node_modules/mammoth/{mammoth.browser.js,mammoth.browser.min.js}"
-  - '!node_modules/selection-hook/prebuilds/**/*' # we rebuild .node, don't use prebuilds
+  - "!node_modules/selection-hook/prebuilds/**/*" # we rebuild .node, don't use prebuilds
-  - '!node_modules/selection-hook/node_modules' # we don't need what in the node_modules dir
+  - "!node_modules/selection-hook/node_modules" # we don't need what in the node_modules dir
-  - '!node_modules/selection-hook/src' # we don't need source files
+  - "!node_modules/selection-hook/src" # we don't need source files
-  - '!node_modules/tesseract.js-core/{tesseract-core.js,tesseract-core.wasm,tesseract-core.wasm.js}' # we don't need source files
+  - "!node_modules/tesseract.js-core/{tesseract-core.js,tesseract-core.wasm,tesseract-core.wasm.js}" # we don't need source files
-  - '!node_modules/tesseract.js-core/{tesseract-core-lstm.js,tesseract-core-lstm.wasm,tesseract-core-lstm.wasm.js}' # we don't need source files
+  - "!node_modules/tesseract.js-core/{tesseract-core-lstm.js,tesseract-core-lstm.wasm,tesseract-core-lstm.wasm.js}" # we don't need source files
-  - '!node_modules/tesseract.js-core/{tesseract-core-simd-lstm.js,tesseract-core-simd-lstm.wasm,tesseract-core-simd-lstm.wasm.js}' # we don't need source files
+  - "!node_modules/tesseract.js-core/{tesseract-core-simd-lstm.js,tesseract-core-simd-lstm.wasm,tesseract-core-simd-lstm.wasm.js}" # we don't need source files
-  - '!**/*.{h,iobj,ipdb,tlog,recipe,vcxproj,vcxproj.filters,Makefile,*.Makefile}' # filter .node build files
+  - "!**/*.{h,iobj,ipdb,tlog,recipe,vcxproj,vcxproj.filters,Makefile,*.Makefile}" # filter .node build files
 asarUnpack:
   - resources/**
-  - '**/*.{metal,exp,lib}'
+  - "**/*.{metal,exp,lib}"
+  - "node_modules/@img/sharp-libvips-*/**"
 win:
   executableName: Cherry Studio
   artifactName: ${productName}-${version}-${arch}-setup.${ext}
@@ -87,7 +88,7 @@ mac:
   entitlementsInherit: build/entitlements.mac.plist
   notarize: false
   artifactName: ${productName}-${version}-${arch}.${ext}
-  minimumSystemVersion: '20.1.0' # 最低支持 macOS 11.0
+  minimumSystemVersion: "20.1.0" # 最低支持 macOS 11.0
   extendInfo:
     - NSCameraUsageDescription: Application requests access to the device's camera.
     - NSMicrophoneUsageDescription: Application requests access to the device's microphone.
@@ -109,34 +110,31 @@ linux:
       StartupWMClass: CherryStudio
   mimeTypes:
     - x-scheme-handler/cherrystudio
+rpm:
+  # Workaround for electron build issue on rpm package:
+  # https://github.com/electron/forge/issues/3594
+  fpm: ["--rpm-rpmbuild-define=_build_id_links none"]
 publish:
   provider: generic
   url: https://releases.cherry-ai.com
 electronDownload:
   mirror: https://npmmirror.com/mirrors/electron/
+beforePack: scripts/before-pack.js
 afterPack: scripts/after-pack.js
 afterSign: scripts/notarize.js
 artifactBuildCompleted: scripts/artifact-build-completed.js
 releaseInfo:
   releaseNotes: |
-    ✨ 重要更新：
+    🐛 问题修复：
-    - 新增笔记模块，支持富文本编辑和管理
+    - 修复 Anthropic API URL 处理，移除尾部斜杠并添加端点路径处理
-    - 内置 GLM-4.5-Flash 免费模型（由智谱开放平台提供）
+    - 修复 MessageEditor 缺少 QuickPanelProvider 包装的问题
-    - 内置 Qwen3-8B 免费模型（由硅基流动提供）
+    - 修复 MiniWindow 高度问题
-    - 新增 Nano Banana（Gemini 2.5 Flash Image）模型支持
-    - 新增系统 OCR 功能 (macOS & Windows)
-    - 新增图片 OCR 识别和翻译功能
-    - 模型切换支持通过标签筛选
-    - 翻译功能增强：历史搜索和收藏
 
-    🔧 性能优化：
+    🚀 性能优化：
-    - 优化历史页面搜索性能
+    - 优化输入栏提及模型状态缓存，在渲染间保持状态
-    - 优化拖拽列表组件交互
+    - 重构网络搜索参数支持模型内置搜索，新增 OpenAI Chat 和 OpenRouter 支持
-    - 升级 Electron 到 37.4.0
 
-    🐛 修复问题：
+    🔧 重构改进：
-    - 修复知识库加密 PDF 文档处理
+    - 更新 HeroUIProvider 导入路径，改善上下文管理
-    - 修复导航栏在左侧时笔记侧边栏按钮缺失
+    - 更新依赖项和 VSCode 开发环境配置
-    - 修复多个模型兼容性问题
+    - 升级 @cherrystudio/ai-core 到 v1.0.0-alpha.17
-    - 修复 MCP 相关问题
-    - 其他稳定性改进

electron.vite.config.ts

@@ -4,6 +4,10 @@ import { defineConfig, externalizeDepsPlugin } from 'electron-vite'
 import { resolve } from 'path'
 import { visualizer } from 'rollup-plugin-visualizer'
 
+// assert not supported by biome
+// import pkg from './package.json' assert { type: 'json' }
+import pkg from './package.json'
+
 const visualizerPlugin = (type: 'renderer' | 'main') => {
   return process.env[`VISUALIZER_${type.toUpperCase()}`] ? [visualizer({ open: true })] : []
 }
@@ -26,7 +30,7 @@ export default defineConfig({
     },
     build: {
       rollupOptions: {
-        external: ['@libsql/client', 'bufferutil', 'utf-8-validate'],
+        external: ['bufferutil', 'utf-8-validate', 'electron', ...Object.keys(pkg.dependencies)],
         output: {
           manualChunks: undefined, // 彻底禁用代码分割 - 返回 null 强制单文件打包
           inlineDynamicImports: true // 内联所有动态导入，这是关键配置
@@ -58,6 +62,7 @@ export default defineConfig({
   },
   renderer: {
     plugins: [
+      (async () => (await import('@tailwindcss/vite')).default())(),
       react({
         tsDecorators: true,
         plugins: [
@@ -79,9 +84,13 @@ export default defineConfig({
       alias: {
         '@renderer': resolve('src/renderer/src'),
         '@shared': resolve('packages/shared'),
+        '@types': resolve('src/renderer/src/types'),
         '@logger': resolve('src/renderer/src/services/LoggerService'),
         '@mcp-trace/trace-core': resolve('packages/mcp-trace/trace-core'),
         '@mcp-trace/trace-web': resolve('packages/mcp-trace/trace-web'),
+        '@cherrystudio/ai-core/provider': resolve('packages/aiCore/src/core/providers'),
+        '@cherrystudio/ai-core/built-in/plugins': resolve('packages/aiCore/src/core/plugins/built-in'),
+        '@cherrystudio/ai-core': resolve('packages/aiCore/src'),
         '@cherrystudio/extension-table-plus': resolve('packages/extension-table-plus/src')
       }
     },

eslint.config.mjs

@@ -1,8 +1,8 @@
-import electronConfigPrettier from '@electron-toolkit/eslint-config-prettier'
 import tseslint from '@electron-toolkit/eslint-config-ts'
 import eslint from '@eslint/js'
 import eslintReact from '@eslint-react/eslint-plugin'
 import { defineConfig } from 'eslint/config'
+import oxlint from 'eslint-plugin-oxlint'
 import reactHooks from 'eslint-plugin-react-hooks'
 import simpleImportSort from 'eslint-plugin-simple-import-sort'
 import unusedImports from 'eslint-plugin-unused-imports'
@@ -10,7 +10,6 @@ import unusedImports from 'eslint-plugin-unused-imports'
 export default defineConfig([
   eslint.configs.recommended,
   tseslint.configs.recommended,
-  electronConfigPrettier,
   eslintReact.configs['recommended-typescript'],
   reactHooks.configs['recommended-latest'],
   {
@@ -26,7 +25,6 @@ export default defineConfig([
       'simple-import-sort/exports': 'error',
       'unused-imports/no-unused-imports': 'error',
       '@eslint-react/no-prop-types': 'error',
-      'prettier/prettier': ['error']
     }
   },
   // Configuration for ensuring compatibility with the original ESLint(8.x) rules
@@ -53,7 +51,7 @@ export default defineConfig([
   {
     // LoggerService Custom Rules - only apply to src directory
     files: ['src/**/*.{ts,tsx,js,jsx}'],
-    ignores: ['src/**/__tests__/**', 'src/**/__mocks__/**', 'src/**/*.test.*'],
+    ignores: ['src/**/__tests__/**', 'src/**/__mocks__/**', 'src/**/*.test.*', 'src/preload/**'],
     rules: {
       'no-restricted-syntax': [
         process.env.PRCI ? 'error' : 'warn',
@@ -123,7 +121,14 @@ export default defineConfig([
       '.gitignore',
       'scripts/cloudflare-worker.js',
       'src/main/integration/nutstore/sso/lib/**',
-      'src/main/integration/cherryin/index.js'
+      'src/main/integration/cherryin/index.js',
+      'src/main/integration/nutstore/sso/lib/**',
+      'src/renderer/src/ui/**',
+      'packages/**/dist'
     ]
-  }
+  },
+  // turn off oxlint supported rules.
+  ...oxlint.configs['flat/eslint'],
+  ...oxlint.configs['flat/typescript'],
+  ...oxlint.configs['flat/unicorn']
 ])

package.json

@@ -1,6 +1,6 @@
 {
   "name": "CherryStudio",
-  "version": "1.5.8",
+  "version": "1.6.0-rc.2",
   "private": true,
   "description": "A powerful AI assistant for producer.",
   "main": "./out/main/index.js",
@@ -27,6 +27,7 @@
   "scripts": {
     "start": "electron-vite preview",
     "dev": "dotenv electron-vite dev",
+    "dev:main": "dotenv electron-vite dev --watch",
     "debug": "electron-vite -- --inspect --sourcemap --remote-debugging-port=9222",
     "build": "npm run typecheck && electron-vite build",
     "build:check": "yarn lint && yarn test",
@@ -40,19 +41,21 @@
     "build:linux": "dotenv npm run build && electron-builder --linux --x64 --arm64",
     "build:linux:arm64": "dotenv npm run build && electron-builder --linux --arm64",
     "build:linux:x64": "dotenv npm run build && electron-builder --linux --x64",
-    "build:npm": "node scripts/build-npm.js",
     "release": "node scripts/version.js",
     "publish": "yarn build:check && yarn release patch push",
     "pulish:artifacts": "cd packages/artifacts && npm publish && cd -",
-    "generate:agents": "yarn workspace @cherry-studio/database agents",
+    "agents:generate": "NODE_ENV='development' drizzle-kit generate --config src/main/services/agents/drizzle.config.ts",
+    "agents:push": "NODE_ENV='development' drizzle-kit push --config src/main/services/agents/drizzle.config.ts",
+    "agents:studio": "NODE_ENV='development' drizzle-kit studio --config src/main/services/agents/drizzle.config.ts",
+    "agents:drop": "NODE_ENV='development' drizzle-kit drop --config src/main/services/agents/drizzle.config.ts",
     "generate:icons": "electron-icon-builder --input=./build/logo.png --output=build",
     "analyze:renderer": "VISUALIZER_RENDERER=true yarn build",
     "analyze:main": "VISUALIZER_MAIN=true yarn build",
-    "typecheck": "npm run typecheck:node && npm run typecheck:web",
+    "typecheck": "concurrently -n \"node,web\" -c \"cyan,magenta\" \"npm run typecheck:node\" \"npm run typecheck:web\"",
-    "typecheck:node": "tsc --noEmit -p tsconfig.node.json --composite false",
+    "typecheck:node": "tsgo --noEmit -p tsconfig.node.json --composite false",
-    "typecheck:web": "tsc --noEmit -p tsconfig.web.json --composite false",
+    "typecheck:web": "tsgo --noEmit -p tsconfig.web.json --composite false",
-    "check:i18n": "tsx scripts/check-i18n.ts",
+    "check:i18n": "dotenv -e .env -- tsx scripts/check-i18n.ts",
-    "sync:i18n": "tsx scripts/sync-i18n.ts",
+    "sync:i18n": "dotenv -e .env -- tsx scripts/sync-i18n.ts",
     "update:i18n": "dotenv -e .env -- tsx scripts/update-i18n.ts",
     "auto:i18n": "dotenv -e .env -- tsx scripts/auto-translate-i18n.ts",
     "update:languages": "tsx scripts/update-languages.ts",
@@ -64,24 +67,37 @@
     "test:ui": "vitest --ui",
     "test:watch": "vitest",
     "test:e2e": "yarn playwright test",
-    "test:lint": "eslint . --ext .js,.jsx,.cjs,.mjs,.ts,.tsx,.cts,.mts",
+    "test:lint": "oxlint --deny-warnings && eslint . --ext .js,.jsx,.cjs,.mjs,.ts,.tsx,.cts,.mts --cache",
     "test:scripts": "vitest scripts",
-    "format": "prettier --write .",
+    "lint": "oxlint --fix && eslint . --ext .js,.jsx,.cjs,.mjs,.ts,.tsx,.cts,.mts --fix --cache && yarn typecheck && yarn check:i18n",
-    "lint": "eslint . --ext .js,.jsx,.cjs,.mjs,.ts,.tsx,.cts,.mts --fix && yarn typecheck && yarn check:i18n",
+    "format": "biome format --write && biome lint --write",
-    "prepare": "git config blame.ignoreRevsFile .git-blame-ignore-revs && husky"
+    "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"
   },
   "dependencies": {
+    "@anthropic-ai/claude-code": "patch:@anthropic-ai/claude-code@npm%3A1.0.118#~/.yarn/patches/@anthropic-ai-claude-code-npm-1.0.118-bbf4e9e59f.patch",
     "@libsql/client": "0.14.0",
     "@libsql/win32-x64-msvc": "^0.4.7",
-    "@napi-rs/system-ocr": "^1.0.2",
+    "@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",
     "@strongtz/win32-arm64-msvc": "^0.4.7",
+    "@types/uuid": "^10.0.0",
+    "drizzle-orm": "^0.44.5",
+    "express": "^5.1.0",
+    "express-validator": "^7.2.1",
+    "font-list": "^2.0.0",
     "graceful-fs": "^4.2.11",
     "jsdom": "26.1.0",
     "node-stream-zip": "^1.15.0",
     "officeparser": "^4.2.0",
     "os-proxy-config": "^1.1.2",
-    "selection-hook": "^1.0.11",
+    "selection-hook": "^1.0.12",
     "sharp": "^0.34.3",
+    "swagger-jsdoc": "^6.2.8",
+    "swagger-ui-express": "^5.0.1",
     "tesseract.js": "patch:tesseract.js@npm%3A6.0.1#~/.yarn/patches/tesseract.js-npm-6.0.1-2562a7e46d.patch",
     "turndown": "7.2.0"
   },
@@ -89,12 +105,18 @@
     "@agentic/exa": "^7.3.3",
     "@agentic/searxng": "^7.3.3",
     "@agentic/tavily": "^7.3.3",
+    "@ai-sdk/amazon-bedrock": "^3.0.21",
+    "@ai-sdk/google-vertex": "^3.0.27",
+    "@ai-sdk/mistral": "^2.0.14",
+    "@ai-sdk/perplexity": "^2.0.9",
     "@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",
     "@aws-sdk/client-bedrock": "^3.840.0",
     "@aws-sdk/client-bedrock-runtime": "^3.840.0",
     "@aws-sdk/client-s3": "^3.840.0",
+    "@biomejs/biome": "2.2.4",
+    "@cherrystudio/ai-core": "workspace:^1.0.0-alpha.18",
     "@cherrystudio/embedjs": "^0.1.31",
     "@cherrystudio/embedjs-libsql": "^0.1.31",
     "@cherrystudio/embedjs-loader-csv": "^0.1.31",
@@ -112,7 +134,6 @@
     "@dnd-kit/modifiers": "^9.0.0",
     "@dnd-kit/sortable": "^10.0.0",
     "@dnd-kit/utilities": "^3.2.2",
-    "@electron-toolkit/eslint-config-prettier": "^3.0.0",
     "@electron-toolkit/eslint-config-ts": "^3.0.0",
     "@electron-toolkit/preload": "^3.0.0",
     "@electron-toolkit/tsconfig": "^1.0.1",
@@ -123,13 +144,14 @@
     "@eslint/js": "^9.22.0",
     "@google/genai": "patch:@google/genai@npm%3A1.0.1#~/.yarn/patches/@google-genai-npm-1.0.1-e26f0f9af7.patch",
     "@hello-pangea/dnd": "^18.0.1",
+    "@heroui/react": "^2.8.3",
     "@kangfenmao/keyv-storage": "^0.1.0",
-    "@langchain/community": "^0.3.36",
+    "@langchain/community": "^0.3.50",
-    "@langchain/ollama": "^0.2.1",
     "@mistralai/mistralai": "^1.7.5",
-    "@modelcontextprotocol/sdk": "^1.17.0",
+    "@modelcontextprotocol/sdk": "^1.17.5",
     "@mozilla/readability": "^0.6.0",
     "@notionhq/client": "^2.2.15",
+    "@openrouter/ai-sdk-provider": "^1.1.2",
     "@opentelemetry/api": "^1.9.0",
     "@opentelemetry/core": "2.0.0",
     "@opentelemetry/exporter-trace-otlp-http": "^0.200.0",
@@ -137,9 +159,11 @@
     "@opentelemetry/sdk-trace-node": "^2.0.0",
     "@opentelemetry/sdk-trace-web": "^2.0.0",
     "@playwright/test": "^1.52.0",
+    "@radix-ui/react-context-menu": "^2.2.16",
     "@reduxjs/toolkit": "^2.2.5",
     "@shikijs/markdown-it": "^3.12.0",
-    "@swc/plugin-styled-components": "^7.1.5",
+    "@swc/plugin-styled-components": "^8.0.4",
+    "@tailwindcss/vite": "^4.1.13",
     "@tanstack/react-query": "^5.85.5",
     "@tanstack/react-virtual": "^3.13.12",
     "@testing-library/dom": "^10.4.0",
@@ -165,20 +189,30 @@
     "@truto/turndown-plugin-gfm": "^1.0.2",
     "@tryfabric/martian": "^1.2.4",
     "@types/cli-progress": "^3",
+    "@types/content-type": "^1.1.9",
+    "@types/cors": "^2.8.19",
+    "@types/diff": "^7",
+    "@types/express": "^5",
     "@types/fs-extra": "^11",
     "@types/he": "^1",
+    "@types/html-to-text": "^9",
     "@types/lodash": "^4.17.5",
     "@types/markdown-it": "^14",
     "@types/md5": "^2.3.5",
+    "@types/mime-types": "^3",
     "@types/node": "^22.17.1",
     "@types/pako": "^1.0.2",
     "@types/react": "^19.0.12",
     "@types/react-dom": "^19.0.4",
     "@types/react-infinite-scroll-component": "^5.0.0",
     "@types/react-transition-group": "^4.4.12",
+    "@types/react-window": "^1",
+    "@types/swagger-jsdoc": "^6",
+    "@types/swagger-ui-express": "^4.1.8",
     "@types/tinycolor2": "^1",
     "@types/turndown": "^5.0.5",
     "@types/word-extractor": "^1",
+    "@typescript/native-preview": "latest",
     "@uiw/codemirror-extensions-langs": "^4.25.1",
     "@uiw/codemirror-themes-all": "^4.25.1",
     "@uiw/react-codemirror": "^4.25.1",
@@ -190,16 +224,21 @@
     "@viz-js/lang-dot": "^1.0.5",
     "@viz-js/viz": "^3.14.0",
     "@xyflow/react": "^12.4.4",
+    "ai": "^5.0.44",
     "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",
     "axios": "^1.7.3",
     "browser-image-compression": "^2.0.2",
     "chardet": "^2.1.0",
+    "check-disk-space": "3.4.0",
+    "cheerio": "^1.1.2",
     "chokidar": "^4.0.3",
     "cli-progress": "^3.12.0",
+    "clsx": "^2.1.1",
     "code-inspector-plugin": "^0.20.14",
     "color": "^5.0.0",
+    "concurrently": "^9.2.1",
     "country-flag-emoji-polyfill": "0.1.8",
     "dayjs": "^1.11.11",
     "dexie": "^4.0.8",
@@ -208,9 +247,11 @@
     "docx": "^9.0.2",
     "dompurify": "^3.2.6",
     "dotenv-cli": "^7.4.2",
+    "drizzle-kit": "^0.31.4",
     "electron": "37.4.0",
     "electron-builder": "26.0.15",
     "electron-devtools-installer": "^3.2.0",
+    "electron-reload": "^2.0.0-alpha.1",
     "electron-store": "^8.2.0",
     "electron-updater": "6.6.4",
     "electron-vite": "4.0.0",
@@ -219,18 +260,21 @@
     "emoji-picker-element": "^1.22.1",
     "epub": "patch:epub@npm%3A1.3.0#~/.yarn/patches/epub-npm-1.3.0-8325494ffe.patch",
     "eslint": "^9.22.0",
+    "eslint-plugin-oxlint": "^1.15.0",
     "eslint-plugin-react-hooks": "^5.2.0",
     "eslint-plugin-simple-import-sort": "^12.1.1",
     "eslint-plugin-unused-imports": "^4.1.4",
     "fast-diff": "^1.3.0",
     "fast-xml-parser": "^5.2.0",
     "fetch-socks": "1.3.2",
+    "framer-motion": "^12.23.12",
     "franc-min": "^6.2.0",
     "fs-extra": "^11.2.0",
     "google-auth-library": "^9.15.1",
     "he": "^1.2.0",
     "html-tags": "^5.1.0",
     "html-to-image": "^1.11.13",
+    "html-to-text": "^9.0.5",
     "htmlparser2": "^10.0.0",
     "husky": "^9.1.7",
     "i18next": "^23.11.5",
@@ -247,15 +291,17 @@
     "markdown-it": "^14.1.0",
     "mermaid": "^11.10.1",
     "mime": "^4.0.4",
+    "mime-types": "^3.0.1",
     "motion": "^12.10.5",
     "notion-helper": "^1.3.22",
     "npx-scope-finder": "^1.2.0",
     "openai": "patch:openai@npm%3A5.12.2#~/.yarn/patches/openai-npm-5.12.2-30b075401c.patch",
+    "oxlint": "^1.15.0",
+    "oxlint-tsgolint": "^0.2.0",
     "p-queue": "^8.1.0",
     "pdf-lib": "^1.17.1",
+    "pdf-parse": "^1.1.1",
     "playwright": "^1.52.0",
-    "prettier": "^3.5.3",
-    "prettier-plugin-sort-json": "^4.1.1",
     "proxy-agent": "^6.5.0",
     "react": "^19.0.0",
     "react-dom": "^19.0.0",
@@ -265,6 +311,7 @@
     "react-infinite-scroll-component": "^6.1.0",
     "react-json-view": "^1.21.3",
     "react-markdown": "^10.1.0",
+    "react-player": "^3.3.1",
     "react-redux": "^9.1.2",
     "react-router": "6",
     "react-router-dom": "6",
@@ -284,21 +331,23 @@
     "remark-math": "^6.0.0",
     "remove-markdown": "^0.6.2",
     "rollup-plugin-visualizer": "^5.12.0",
-    "sass": "^1.88.0",
     "shiki": "^3.12.0",
     "strict-url-sanitise": "^0.0.1",
     "string-width": "^7.2.0",
     "striptags": "^3.2.0",
     "styled-components": "^6.1.11",
+    "swr": "^2.3.6",
+    "tailwindcss": "^4.1.13",
     "tar": "^7.4.3",
     "tiny-pinyin": "^1.3.2",
     "tokenx": "^1.1.0",
     "tsx": "^4.20.3",
     "turndown-plugin-gfm": "^1.0.2",
-    "typescript": "^5.6.2",
+    "tw-animate-css": "^1.3.8",
+    "typescript": "~5.8.2",
     "undici": "6.21.2",
     "unified": "^11.0.5",
-    "uuid": "^10.0.0",
+    "uuid": "^13.0.0",
     "vite": "npm:rolldown-vite@latest",
     "vitest": "^3.2.4",
     "webdav": "^5.8.0",
@@ -306,9 +355,11 @@
     "winston-daily-rotate-file": "^5.0.0",
     "word-extractor": "^1.0.4",
     "y-protocols": "^1.0.6",
+    "yaml": "^2.8.1",
     "yjs": "^13.6.27",
+    "youtubei.js": "^15.0.1",
     "zipread": "^1.3.3",
-    "zod": "^3.25.74"
+    "zod": "^4.1.5"
   },
   "resolutions": {
     "@codemirror/language": "6.11.3",
@@ -334,11 +385,11 @@
   "packageManager": "yarn@4.9.1",
   "lint-staged": {
     "*.{js,jsx,ts,tsx,cjs,mjs,cts,mts}": [
-      "prettier --write",
+      "biome format --write --no-errors-on-unmatched",
       "eslint --fix"
     ],
-    "*.{json,md,yml,yaml,css,scss,html}": [
+    "*.{json,yml,yaml,css,html}": [
-      "prettier --write"
+      "biome format --write --no-errors-on-unmatched"
     ]
   }
 }

packages/aiCore/AI_SDK_ARCHITECTURE.md

@@ -0,0 +1,514 @@
+# AI Core 基于 Vercel AI SDK 的技术架构
+
+## 1. 架构设计理念
+
+### 1.1 设计目标
+
+- **简化分层**：`models`（模型层）→ `runtime`（运行时层），清晰的职责分离
+- **统一接口**：使用 Vercel AI SDK 统一不同 AI Provider 的接口差异
+- **动态导入**：通过动态导入实现按需加载，减少打包体积
+- **最小包装**：直接使用 AI SDK 的类型和接口，避免重复定义
+- **插件系统**：基于钩子的通用插件架构，支持请求全生命周期扩展
+- **类型安全**：利用 TypeScript 和 AI SDK 的类型系统确保类型安全
+- **轻量级**：专注核心功能，保持包的轻量和高效
+- **包级独立**：作为独立包管理，便于复用和维护
+- **Agent就绪**：为将来集成 OpenAI Agents SDK 预留扩展空间
+
+### 1.2 核心优势
+
+- **标准化**：AI SDK 提供统一的模型接口，减少适配工作
+- **简化设计**：函数式API，避免过度抽象
+- **更好的开发体验**：完整的 TypeScript 支持和丰富的生态系统
+- **性能优化**：AI SDK 内置优化和最佳实践
+- **模块化设计**：独立包结构，支持跨项目复用
+- **可扩展插件**：通用的流转换和参数处理插件系统
+- **面向未来**：为 OpenAI Agents SDK 集成做好准备
+
+## 2. 整体架构图
+
+```mermaid
+graph TD
+    subgraph "用户应用 (如 Cherry Studio)"
+        UI["用户界面"]
+        Components["应用组件"]
+    end
+
+    subgraph "packages/aiCore (AI Core 包)"
+        subgraph "Runtime Layer (运行时层)"
+            RuntimeExecutor["RuntimeExecutor (运行时执行器)"]
+            PluginEngine["PluginEngine (插件引擎)"]
+            RuntimeAPI["Runtime API (便捷函数)"]
+        end
+
+        subgraph "Models Layer (模型层)"
+            ModelFactory["createModel() (模型工厂)"]
+            ProviderCreator["ProviderCreator (提供商创建器)"]
+        end
+
+        subgraph "Core Systems (核心系统)"
+            subgraph "Plugins (插件)"
+                PluginManager["PluginManager (插件管理)"]
+                BuiltInPlugins["Built-in Plugins (内置插件)"]
+                StreamTransforms["Stream Transforms (流转换)"]
+            end
+
+            subgraph "Middleware (中间件)"
+                MiddlewareWrapper["wrapModelWithMiddlewares() (中间件包装)"]
+            end
+
+            subgraph "Providers (提供商)"
+                Registry["Provider Registry (注册表)"]
+                Factory["Provider Factory (工厂)"]
+            end
+        end
+    end
+
+    subgraph "Vercel AI SDK"
+        AICore["ai (核心库)"]
+        OpenAI["@ai-sdk/openai"]
+        Anthropic["@ai-sdk/anthropic"]
+        Google["@ai-sdk/google"]
+        XAI["@ai-sdk/xai"]
+        Others["其他 19+ Providers"]
+    end
+
+    subgraph "Future: OpenAI Agents SDK"
+        AgentSDK["@openai/agents (未来集成)"]
+        AgentExtensions["Agent Extensions (预留)"]
+    end
+
+    UI --> RuntimeAPI
+    Components --> RuntimeExecutor
+    RuntimeAPI --> RuntimeExecutor
+    RuntimeExecutor --> PluginEngine
+    RuntimeExecutor --> ModelFactory
+    PluginEngine --> PluginManager
+    ModelFactory --> ProviderCreator
+    ModelFactory --> MiddlewareWrapper
+    ProviderCreator --> Registry
+    Registry --> Factory
+    Factory --> OpenAI
+    Factory --> Anthropic
+    Factory --> Google
+    Factory --> XAI
+    Factory --> Others
+
+    RuntimeExecutor --> AICore
+    AICore --> streamText
+    AICore --> generateText
+    AICore --> streamObject
+    AICore --> generateObject
+
+    PluginManager --> StreamTransforms
+    PluginManager --> BuiltInPlugins
+
+    %% 未来集成路径
+    RuntimeExecutor -.-> AgentSDK
+    AgentSDK -.-> AgentExtensions
+```
+
+## 3. 包结构设计
+
+### 3.1 新架构文件结构
+
+```
+packages/aiCore/
+├── src/
+│   ├── core/                        # 核心层 - 内部实现
+│   │   ├── models/                  # 模型层 - 模型创建和配置
+│   │   │   ├── factory.ts           # 模型工厂函数 ✅
+│   │   │   ├── ModelCreator.ts      # 模型创建器 ✅
+│   │   │   ├── ConfigManager.ts     # 配置管理器 ✅
+│   │   │   ├── types.ts             # 模型类型定义 ✅
+│   │   │   └── index.ts             # 模型层导出 ✅
+│   │   ├── runtime/                 # 运行时层 - 执行和用户API
+│   │   │   ├── executor.ts          # 运行时执行器 ✅
+│   │   │   ├── pluginEngine.ts      # 插件引擎 ✅
+│   │   │   ├── types.ts             # 运行时类型定义 ✅
+│   │   │   └── index.ts             # 运行时导出 ✅
+│   │   ├── middleware/              # 中间件系统
+│   │   │   ├── wrapper.ts           # 模型包装器 ✅
+│   │   │   ├── manager.ts           # 中间件管理器 ✅
+│   │   │   ├── types.ts             # 中间件类型 ✅
+│   │   │   └── index.ts             # 中间件导出 ✅
+│   │   ├── plugins/                 # 插件系统
+│   │   │   ├── types.ts             # 插件类型定义 ✅
+│   │   │   ├── manager.ts           # 插件管理器 ✅
+│   │   │   ├── built-in/            # 内置插件 ✅
+│   │   │   │   ├── logging.ts       # 日志插件 ✅
+│   │   │   │   ├── webSearchPlugin/ # 网络搜索插件 ✅
+│   │   │   │   ├── toolUsePlugin/   # 工具使用插件 ✅
+│   │   │   │   └── index.ts         # 内置插件导出 ✅
+│   │   │   ├── README.md            # 插件文档 ✅
+│   │   │   └── index.ts             # 插件导出 ✅
+│   │   ├── providers/               # 提供商管理
+│   │   │   ├── registry.ts          # 提供商注册表 ✅
+│   │   │   ├── factory.ts           # 提供商工厂 ✅
+│   │   │   ├── creator.ts           # 提供商创建器 ✅
+│   │   │   ├── types.ts             # 提供商类型 ✅
+│   │   │   ├── utils.ts             # 工具函数 ✅
+│   │   │   └── index.ts             # 提供商导出 ✅
+│   │   ├── options/                 # 配置选项
+│   │   │   ├── factory.ts           # 选项工厂 ✅
+│   │   │   ├── types.ts             # 选项类型 ✅
+│   │   │   ├── xai.ts               # xAI 选项 ✅
+│   │   │   ├── openrouter.ts        # OpenRouter 选项 ✅
+│   │   │   ├── examples.ts          # 示例配置 ✅
+│   │   │   └── index.ts             # 选项导出 ✅
+│   │   └── index.ts                 # 核心层导出 ✅
+│   ├── types.ts                     # 全局类型定义 ✅
+│   └── index.ts                     # 包主入口文件 ✅
+├── package.json                     # 包配置文件 ✅
+├── tsconfig.json                    # TypeScript 配置 ✅
+├── README.md                        # 包说明文档 ✅
+└── AI_SDK_ARCHITECTURE.md           # 本文档 ✅
+```
+
+## 4. 架构分层详解
+
+### 4.1 Models Layer (模型层)
+
+**职责**：统一的模型创建和配置管理
+
+**核心文件**：
+
+- `factory.ts`: 模型工厂函数 (`createModel`, `createModels`)
+- `ProviderCreator.ts`: 底层提供商创建和模型实例化
+- `types.ts`: 模型配置类型定义
+
+**设计特点**：
+
+- 函数式设计，避免不必要的类抽象
+- 统一的模型配置接口
+- 自动处理中间件应用
+- 支持批量模型创建
+
+**核心API**：
+
+```typescript
+// 模型配置接口
+export interface ModelConfig {
+  providerId: ProviderId
+  modelId: string
+  options: ProviderSettingsMap[ProviderId]
+  middlewares?: LanguageModelV1Middleware[]
+}
+
+// 核心模型创建函数
+export async function createModel(config: ModelConfig): Promise<LanguageModel>
+export async function createModels(configs: ModelConfig[]): Promise<LanguageModel[]>
+```
+
+### 4.2 Runtime Layer (运行时层)
+
+**职责**：运行时执行器和用户面向的API接口
+
+**核心组件**：
+
+- `executor.ts`: 运行时执行器类
+- `plugin-engine.ts`: 插件引擎（原PluginEnabledAiClient）
+- `index.ts`: 便捷函数和工厂方法
+
+**设计特点**：
+
+- 提供三种使用方式：类实例、静态工厂、函数式调用
+- 自动集成模型创建和插件处理
+- 完整的类型安全支持
+- 为 OpenAI Agents SDK 预留扩展接口
+
+**核心API**：
+
+```typescript
+// 运行时执行器
+export class RuntimeExecutor<T extends ProviderId = ProviderId> {
+  static create<T extends ProviderId>(
+    providerId: T,
+    options: ProviderSettingsMap[T],
+    plugins?: AiPlugin[]
+  ): RuntimeExecutor<T>
+
+  async streamText(modelId: string, params: StreamTextParams): Promise<StreamTextResult>
+  async generateText(modelId: string, params: GenerateTextParams): Promise<GenerateTextResult>
+  async streamObject(modelId: string, params: StreamObjectParams): Promise<StreamObjectResult>
+  async generateObject(modelId: string, params: GenerateObjectParams): Promise<GenerateObjectResult>
+}
+
+// 便捷函数式API
+export async function streamText<T extends ProviderId>(
+  providerId: T,
+  options: ProviderSettingsMap[T],
+  modelId: string,
+  params: StreamTextParams,
+  plugins?: AiPlugin[]
+): Promise<StreamTextResult>
+```
+
+### 4.3 Plugin System (插件系统)
+
+**职责**：可扩展的插件架构
+
+**核心组件**：
+
+- `PluginManager`: 插件生命周期管理
+- `built-in/`: 内置插件集合
+- 流转换收集和应用
+
+**设计特点**：
+
+- 借鉴 Rollup 的钩子分类设计
+- 支持流转换 (`experimental_transform`)
+- 内置常用插件（日志、计数等）
+- 完整的生命周期钩子
+
+**插件接口**：
+
+```typescript
+export interface AiPlugin {
+  name: string
+  enforce?: 'pre' | 'post'
+
+  // 【First】首个钩子 - 只执行第一个返回值的插件
+  resolveModel?: (modelId: string, context: AiRequestContext) => string | null | Promise<string | null>
+  loadTemplate?: (templateName: string, context: AiRequestContext) => any | null | Promise<any | null>
+
+  // 【Sequential】串行钩子 - 链式执行，支持数据转换
+  transformParams?: (params: any, context: AiRequestContext) => any | Promise<any>
+  transformResult?: (result: any, context: AiRequestContext) => any | Promise<any>
+
+  // 【Parallel】并行钩子 - 不依赖顺序，用于副作用
+  onRequestStart?: (context: AiRequestContext) => void | Promise<void>
+  onRequestEnd?: (context: AiRequestContext, result: any) => void | Promise<void>
+  onError?: (error: Error, context: AiRequestContext) => void | Promise<void>
+
+  // 【Stream】流处理
+  transformStream?: () => TransformStream
+}
+```
+
+### 4.4 Middleware System (中间件系统)
+
+**职责**：AI SDK原生中间件支持
+
+**核心组件**：
+
+- `ModelWrapper.ts`: 模型包装函数
+
+**设计哲学**：
+
+- 直接使用AI SDK的 `wrapLanguageModel`
+- 与插件系统分离，职责明确
+- 函数式设计，简化使用
+
+```typescript
+export function wrapModelWithMiddlewares(model: LanguageModel, middlewares: LanguageModelV1Middleware[]): LanguageModel
+```
+
+### 4.5 Provider System (提供商系统)
+
+**职责**：AI Provider注册表和动态导入
+
+**核心组件**：
+
+- `registry.ts`: 19+ Provider配置和类型
+- `factory.ts`: Provider配置工厂
+
+**支持的Providers**：
+
+- OpenAI, Anthropic, Google, XAI
+- Azure OpenAI, Amazon Bedrock, Google Vertex
+- Groq, Together.ai, Fireworks, DeepSeek
+- 等19+ AI SDK官方支持的providers
+
+## 5. 使用方式
+
+### 5.1 函数式调用 (推荐 - 简单场景)
+
+```typescript
+import { streamText, generateText } from '@cherrystudio/ai-core/runtime'
+
+// 直接函数调用
+const stream = await streamText(
+  'anthropic',
+  { apiKey: 'your-api-key' },
+  'claude-3',
+  { messages: [{ role: 'user', content: 'Hello!' }] },
+  [loggingPlugin]
+)
+```
+
+### 5.2 执行器实例 (推荐 - 复杂场景)
+
+```typescript
+import { createExecutor } from '@cherrystudio/ai-core/runtime'
+
+// 创建可复用的执行器
+const executor = createExecutor('openai', { apiKey: 'your-api-key' }, [plugin1, plugin2])
+
+// 多次使用
+const stream = await executor.streamText('gpt-4', {
+  messages: [{ role: 'user', content: 'Hello!' }]
+})
+
+const result = await executor.generateText('gpt-4', {
+  messages: [{ role: 'user', content: 'How are you?' }]
+})
+```
+
+### 5.3 静态工厂方法
+
+```typescript
+import { RuntimeExecutor } from '@cherrystudio/ai-core/runtime'
+
+// 静态创建
+const executor = RuntimeExecutor.create('anthropic', { apiKey: 'your-api-key' })
+await executor.streamText('claude-3', { messages: [...] })
+```
+
+### 5.4 直接模型创建 (高级用法)
+
+```typescript
+import { createModel } from '@cherrystudio/ai-core/models'
+import { streamText } from 'ai'
+
+// 直接创建模型使用
+const model = await createModel({
+  providerId: 'openai',
+  modelId: 'gpt-4',
+  options: { apiKey: 'your-api-key' },
+  middlewares: [middleware1, middleware2]
+})
+
+// 直接使用 AI SDK
+const result = await streamText({ model, messages: [...] })
+```
+
+## 6. 为 OpenAI Agents SDK 预留的设计
+
+### 6.1 架构兼容性
+
+当前架构完全兼容 OpenAI Agents SDK 的集成需求：
+
+```typescript
+// 当前的模型创建
+const model = await createModel({
+  providerId: 'anthropic',
+  modelId: 'claude-3',
+  options: { apiKey: 'xxx' }
+})
+
+// 将来可以直接用于 OpenAI Agents SDK
+import { Agent, run } from '@openai/agents'
+
+const agent = new Agent({
+  model, // ✅ 直接兼容 LanguageModel 接口
+  name: 'Assistant',
+  instructions: '...',
+  tools: [tool1, tool2]
+})
+
+const result = await run(agent, 'user input')
+```
+
+### 6.2 预留的扩展点
+
+1. **runtime/agents/** 目录预留
+2. **AgentExecutor** 类预留
+3. **Agent工具转换插件** 预留
+4. **多Agent编排** 预留
+
+### 6.3 未来架构扩展
+
+```
+packages/aiCore/src/core/
+├── runtime/
+│   ├── agents/              # 🚀 未来添加
+│   │   ├── AgentExecutor.ts
+│   │   ├── WorkflowManager.ts
+│   │   └── ConversationManager.ts
+│   ├── executor.ts
+│   └── index.ts
+```
+
+## 7. 架构优势
+
+### 7.1 简化设计
+
+- **移除过度抽象**：删除了orchestration层和creation层的复杂包装
+- **函数式优先**：models层使用函数而非类
+- **直接明了**：runtime层直接提供用户API
+
+### 7.2 职责清晰
+
+- **Models**: 专注模型创建和配置
+- **Runtime**: 专注执行和用户API
+- **Plugins**: 专注扩展功能
+- **Providers**: 专注AI Provider管理
+
+### 7.3 类型安全
+
+- 完整的 TypeScript 支持
+- AI SDK 类型的直接复用
+- 避免类型重复定义
+
+### 7.4 灵活使用
+
+- 三种使用模式满足不同需求
+- 从简单函数调用到复杂执行器
+- 支持直接AI SDK使用
+
+### 7.5 面向未来
+
+- 为 OpenAI Agents SDK 集成做好准备
+- 清晰的扩展点和架构边界
+- 模块化设计便于功能添加
+
+## 8. 技术决策记录
+
+### 8.1 为什么选择简化的两层架构？
+
+- **职责分离**：models专注创建，runtime专注执行
+- **模块化**：每层都有清晰的边界和职责
+- **扩展性**：为Agent功能预留了清晰的扩展空间
+
+### 8.2 为什么选择函数式设计？
+
+- **简洁性**：避免不必要的类设计
+- **性能**：减少对象创建开销
+- **易用性**：函数调用更直观
+
+### 8.3 为什么分离插件和中间件？
+
+- **职责明确**: 插件处理应用特定需求
+- **原生支持**: 中间件使用AI SDK原生功能
+- **灵活性**: 两套系统可以独立演进
+
+## 9. 总结
+
+AI Core架构实现了：
+
+### 9.1 核心特点
+
+- ✅ **简化架构**: 2层核心架构，职责清晰
+- ✅ **函数式设计**: models层完全函数化
+- ✅ **类型安全**: 统一的类型定义和AI SDK类型复用
+- ✅ **插件扩展**: 强大的插件系统
+- ✅ **多种使用方式**: 满足不同复杂度需求
+- ✅ **Agent就绪**: 为OpenAI Agents SDK集成做好准备
+
+### 9.2 核心价值
+
+- **统一接口**: 一套API支持19+ AI providers
+- **灵活使用**: 函数式、实例式、静态工厂式
+- **强类型**: 完整的TypeScript支持
+- **可扩展**: 插件和中间件双重扩展能力
+- **高性能**: 最小化包装，直接使用AI SDK
+- **面向未来**: Agent SDK集成架构就绪
+
+### 9.3 未来发展
+
+这个架构提供了：
+
+- **优秀的开发体验**: 简洁的API和清晰的使用模式
+- **强大的扩展能力**: 为Agent功能预留了完整的架构空间
+- **良好的维护性**: 职责分离明确，代码易于维护
+- **广泛的适用性**: 既适合简单调用也适合复杂应用

packages/aiCore/README.md

@@ -0,0 +1,433 @@
+# @cherrystudio/ai-core
+
+Cherry Studio AI Core 是一个基于 Vercel AI SDK 的统一 AI Provider 接口包，为 AI 应用提供强大的抽象层和插件化架构。
+
+## ✨ 核心亮点
+
+### 🏗️ 优雅的架构设计
+
+- **简化分层**：`models`（模型层）→ `runtime`（运行时层），清晰的职责分离
+- **函数式优先**：避免过度抽象，提供简洁直观的 API
+- **类型安全**：完整的 TypeScript 支持，直接复用 AI SDK 类型系统
+- **最小包装**：直接使用 AI SDK 的接口，避免重复定义和性能损耗
+
+### 🔌 强大的插件系统
+
+- **生命周期钩子**：支持请求全生命周期的扩展点
+- **流转换支持**：基于 AI SDK 的 `experimental_transform` 实现流处理
+- **插件分类**：First、Sequential、Parallel 三种钩子类型，满足不同场景
+- **内置插件**：webSearch、logging、toolUse 等开箱即用的功能
+
+### 🌐 统一多 Provider 接口
+
+- **扩展注册**：支持自定义 Provider 注册，无限扩展能力
+- **配置统一**：统一的配置接口，简化多 Provider 管理
+
+### 🚀 多种使用方式
+
+- **函数式调用**：适合简单场景的直接函数调用
+- **执行器实例**：适合复杂场景的可复用执行器
+- **静态工厂**：便捷的静态创建方法
+- **原生兼容**：完全兼容 AI SDK 原生 Provider Registry
+
+### 🔮 面向未来
+
+- **Agent 就绪**：为 OpenAI Agents SDK 集成预留架构空间
+- **模块化设计**：独立包结构，支持跨项目复用
+- **渐进式迁移**：可以逐步从现有 AI SDK 代码迁移
+
+## 特性
+
+- 🚀 统一的 AI Provider 接口
+- 🔄 动态导入支持
+- 🛠️ TypeScript 支持
+- 📦 强大的插件系统
+- 🌍 内置webSearch(Openai,Google,Anthropic,xAI)
+- 🎯 多种使用模式（函数式/实例式/静态工厂）
+- 🔌 可扩展的 Provider 注册系统
+- 🧩 完整的中间件支持
+- 📊 插件统计和调试功能
+
+## 支持的 Providers
+
+基于 [AI SDK 官方支持的 providers](https://ai-sdk.dev/providers/ai-sdk-providers)：
+
+**核心 Providers（内置支持）:**
+
+- OpenAI
+- Anthropic
+- Google Generative AI
+- OpenAI-Compatible
+- xAI (Grok)
+- Azure OpenAI
+- DeepSeek
+
+**扩展 Providers（通过注册API支持）:**
+
+- Google Vertex AI
+- ...
+- 自定义 Provider
+
+## 安装
+
+```bash
+npm install @cherrystudio/ai-core ai
+```
+
+### React Native
+
+如果你在 React Native 项目中使用此包，需要在 `metro.config.js` 中添加以下配置：
+
+```javascript
+// metro.config.js
+const { getDefaultConfig } = require('expo/metro-config')
+
+const config = getDefaultConfig(__dirname)
+
+// 添加对 @cherrystudio/ai-core 的支持
+config.resolver.resolverMainFields = ['react-native', 'browser', 'main']
+config.resolver.platforms = ['ios', 'android', 'native', 'web']
+
+module.exports = config
+```
+
+还需要安装你要使用的 AI SDK provider:
+
+```bash
+npm install @ai-sdk/openai @ai-sdk/anthropic @ai-sdk/google
+```
+
+## 使用示例
+
+### 基础用法
+
+```typescript
+import { AiCore } from '@cherrystudio/ai-core'
+
+// 创建 OpenAI executor
+const executor = AiCore.create('openai', {
+  apiKey: 'your-api-key'
+})
+
+// 流式生成
+const result = await executor.streamText('gpt-4', {
+  messages: [{ role: 'user', content: 'Hello!' }]
+})
+
+// 非流式生成
+const response = await executor.generateText('gpt-4', {
+  messages: [{ role: 'user', content: 'Hello!' }]
+})
+```
+
+### 便捷函数
+
+```typescript
+import { createOpenAIExecutor } from '@cherrystudio/ai-core'
+
+// 快速创建 OpenAI executor
+const executor = createOpenAIExecutor({
+  apiKey: 'your-api-key'
+})
+
+// 使用 executor
+const result = await executor.streamText('gpt-4', {
+  messages: [{ role: 'user', content: 'Hello!' }]
+})
+```
+
+### 多 Provider 支持
+
+```typescript
+import { AiCore } from '@cherrystudio/ai-core'
+
+// 支持多种 AI providers
+const openaiExecutor = AiCore.create('openai', { apiKey: 'openai-key' })
+const anthropicExecutor = AiCore.create('anthropic', { apiKey: 'anthropic-key' })
+const googleExecutor = AiCore.create('google', { apiKey: 'google-key' })
+const xaiExecutor = AiCore.create('xai', { apiKey: 'xai-key' })
+```
+
+### 扩展 Provider 注册
+
+对于非内置的 providers，可以通过注册 API 扩展支持：
+
+```typescript
+import { registerProvider, AiCore } from '@cherrystudio/ai-core'
+
+// 方式一：导入并注册第三方 provider
+import { createGroq } from '@ai-sdk/groq'
+
+registerProvider({
+  id: 'groq',
+  name: 'Groq',
+  creator: createGroq,
+  supportsImageGeneration: false
+})
+
+// 现在可以使用 Groq
+const groqExecutor = AiCore.create('groq', { apiKey: 'groq-key' })
+
+// 方式二：动态导入方式注册
+registerProvider({
+  id: 'mistral',
+  name: 'Mistral AI',
+  import: () => import('@ai-sdk/mistral'),
+  creatorFunctionName: 'createMistral'
+})
+
+const mistralExecutor = AiCore.create('mistral', { apiKey: 'mistral-key' })
+```
+
+## 🔌 插件系统
+
+AI Core 提供了强大的插件系统，支持请求全生命周期的扩展。
+
+### 内置插件
+
+#### webSearchPlugin - 网络搜索插件
+
+为不同 AI Provider 提供统一的网络搜索能力：
+
+```typescript
+import { webSearchPlugin } from '@cherrystudio/ai-core/built-in/plugins'
+
+const executor = AiCore.create('openai', { apiKey: 'your-key' }, [
+  webSearchPlugin({
+    openai: {
+      /* OpenAI 搜索配置 */
+    },
+    anthropic: { maxUses: 5 },
+    google: {
+      /* Google 搜索配置 */
+    },
+    xai: {
+      mode: 'on',
+      returnCitations: true,
+      maxSearchResults: 5,
+      sources: [{ type: 'web' }, { type: 'x' }, { type: 'news' }]
+    }
+  })
+])
+```
+
+#### loggingPlugin - 日志插件
+
+提供详细的请求日志记录：
+
+```typescript
+import { createLoggingPlugin } from '@cherrystudio/ai-core/built-in/plugins'
+
+const executor = AiCore.create('openai', { apiKey: 'your-key' }, [
+  createLoggingPlugin({
+    logLevel: 'info',
+    includeParams: true,
+    includeResult: false
+  })
+])
+```
+
+#### promptToolUsePlugin - 提示工具使用插件
+
+为不支持原生 Function Call 的模型提供 prompt 方式的工具调用：
+
+```typescript
+import { createPromptToolUsePlugin } from '@cherrystudio/ai-core/built-in/plugins'
+
+// 对于不支持 function call 的模型
+const executor = AiCore.create(
+  'providerId',
+  {
+    apiKey: 'your-key',
+    baseURL: 'https://your-model-endpoint'
+  },
+  [
+    createPromptToolUsePlugin({
+      enabled: true,
+      // 可选：自定义系统提示符构建
+      buildSystemPrompt: (userPrompt, tools) => {
+        return `${userPrompt}\n\nAvailable tools: ${Object.keys(tools).join(', ')}`
+      }
+    })
+  ]
+)
+```
+
+### 自定义插件
+
+创建自定义插件非常简单：
+
+```typescript
+import { definePlugin } from '@cherrystudio/ai-core'
+
+const customPlugin = definePlugin({
+  name: 'custom-plugin',
+  enforce: 'pre', // 'pre' | 'post' | undefined
+
+  // 在请求开始时记录日志
+  onRequestStart: async (context) => {
+    console.log(`Starting request for model: ${context.modelId}`)
+  },
+
+  // 转换请求参数
+  transformParams: async (params, context) => {
+    // 添加自定义系统消息
+    if (params.messages) {
+      params.messages.unshift({
+        role: 'system',
+        content: 'You are a helpful assistant.'
+      })
+    }
+    return params
+  },
+
+  // 处理响应结果
+  transformResult: async (result, context) => {
+    // 添加元数据
+    if (result.text) {
+      result.metadata = {
+        processedAt: new Date().toISOString(),
+        modelId: context.modelId
+      }
+    }
+    return result
+  }
+})
+
+// 使用自定义插件
+const executor = AiCore.create('openai', { apiKey: 'your-key' }, [customPlugin])
+```
+
+### 使用 AI SDK 原生 Provider 注册表
+
+> https://ai-sdk.dev/docs/reference/ai-sdk-core/provider-registry
+
+除了使用内建的 provider 管理，你还可以使用 AI SDK 原生的 `createProviderRegistry` 来构建自己的 provider 注册表。
+
+#### 基本用法示例
+
+```typescript
+import { createClient } from '@cherrystudio/ai-core'
+import { createProviderRegistry } from 'ai'
+import { createOpenAI } from '@ai-sdk/openai'
+import { anthropic } from '@ai-sdk/anthropic'
+
+// 1. 创建 AI SDK 原生注册表
+export const registry = createProviderRegistry({
+  // register provider with prefix and default setup:
+  anthropic,
+
+  // register provider with prefix and custom setup:
+  openai: createOpenAI({
+    apiKey: process.env.OPENAI_API_KEY
+  })
+})
+
+// 2. 创建client,'openai'可以传空或者传providerId(内建的provider)
+const client = PluginEnabledAiClient.create('openai', {
+  apiKey: process.env.OPENAI_API_KEY
+})
+
+// 3. 方式1：使用内建逻辑（传统方式）
+const result1 = await client.streamText('gpt-4', {
+  messages: [{ role: 'user', content: 'Hello with built-in logic!' }]
+})
+
+// 4. 方式2：使用自定义注册表（灵活方式）
+const result2 = await client.streamText({
+  model: registry.languageModel('openai:gpt-4'),
+  messages: [{ role: 'user', content: 'Hello with custom registry!' }]
+})
+
+// 5. 支持的重载方法
+await client.generateObject({
+  model: registry.languageModel('openai:gpt-4'),
+  schema: z.object({ name: z.string() }),
+  messages: [{ role: 'user', content: 'Generate a user' }]
+})
+
+await client.streamObject({
+  model: registry.languageModel('anthropic:claude-3-opus-20240229'),
+  schema: z.object({ items: z.array(z.string()) }),
+  messages: [{ role: 'user', content: 'Generate a list' }]
+})
+```
+
+#### 与插件系统配合使用
+
+更强大的是，你还可以将自定义注册表与 Cherry Studio 的插件系统结合使用：
+
+```typescript
+import { PluginEnabledAiClient } from '@cherrystudio/ai-core'
+import { createProviderRegistry } from 'ai'
+import { createOpenAI } from '@ai-sdk/openai'
+import { anthropic } from '@ai-sdk/anthropic'
+
+// 1. 创建带插件的客户端
+const client = PluginEnabledAiClient.create(
+  'openai',
+  {
+    apiKey: process.env.OPENAI_API_KEY
+  },
+  [LoggingPlugin, RetryPlugin]
+)
+
+// 2. 创建自定义注册表
+const registry = createProviderRegistry({
+  openai: createOpenAI({ apiKey: process.env.OPENAI_API_KEY }),
+  anthropic: anthropic({ apiKey: process.env.ANTHROPIC_API_KEY })
+})
+
+// 3. 方式1：使用内建逻辑 + 完整插件系统
+await client.streamText('gpt-4', {
+  messages: [{ role: 'user', content: 'Hello with plugins!' }]
+})
+
+// 4. 方式2：使用自定义注册表 + 有限插件支持
+await client.streamText({
+  model: registry.languageModel('anthropic:claude-3-opus-20240229'),
+  messages: [{ role: 'user', content: 'Hello from Claude!' }]
+})
+
+// 5. 支持的方法
+await client.generateObject({
+  model: registry.languageModel('openai:gpt-4'),
+  schema: z.object({ name: z.string() }),
+  messages: [{ role: 'user', content: 'Generate a user' }]
+})
+
+await client.streamObject({
+  model: registry.languageModel('openai:gpt-4'),
+  schema: z.object({ items: z.array(z.string()) }),
+  messages: [{ role: 'user', content: 'Generate a list' }]
+})
+```
+
+#### 混合使用的优势
+
+- **灵活性**：可以根据需要选择使用内建逻辑或自定义注册表
+- **兼容性**：完全兼容 AI SDK 的 `createProviderRegistry` API
+- **渐进式**：可以逐步迁移现有代码，无需一次性重构
+- **插件支持**：自定义注册表仍可享受插件系统的部分功能
+- **最佳实践**：结合两种方式的优点，既有动态加载的性能优势，又有统一注册表的便利性
+
+## 📚 相关资源
+
+- [Vercel AI SDK 文档](https://ai-sdk.dev/)
+- [Cherry Studio 项目](https://github.com/CherryHQ/cherry-studio)
+- [AI SDK Providers](https://ai-sdk.dev/providers/ai-sdk-providers)
+
+## 未来版本
+
+- 🔮 多 Agent 编排
+- 🔮 可视化插件配置
+- 🔮 实时监控和分析
+- 🔮 云端插件同步
+
+## 📄 License
+
+MIT License - 详见 [LICENSE](https://github.com/CherryHQ/cherry-studio/blob/main/LICENSE) 文件
+
+---
+
+**Cherry Studio AI Core** - 让 AI 开发更简单、更强大、更灵活 🚀

packages/aiCore/package.json

@@ -0,0 +1,85 @@
+{
+  "name": "@cherrystudio/ai-core",
+  "version": "1.0.0-alpha.18",
+  "description": "Cherry Studio AI Core - Unified AI Provider Interface Based on Vercel AI SDK",
+  "main": "dist/index.js",
+  "module": "dist/index.mjs",
+  "types": "dist/index.d.ts",
+  "react-native": "dist/index.js",
+  "scripts": {
+    "build": "tsdown",
+    "dev": "tsc -w",
+    "clean": "rm -rf dist",
+    "test": "vitest run",
+    "test:watch": "vitest"
+  },
+  "keywords": [
+    "ai",
+    "sdk",
+    "openai",
+    "anthropic",
+    "google",
+    "cherry-studio",
+    "vercel-ai-sdk"
+  ],
+  "author": "Cherry Studio",
+  "license": "MIT",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/CherryHQ/cherry-studio.git"
+  },
+  "bugs": {
+    "url": "https://github.com/CherryHQ/cherry-studio/issues"
+  },
+  "homepage": "https://github.com/CherryHQ/cherry-studio#readme",
+  "peerDependencies": {
+    "ai": "^5.0.26"
+  },
+  "dependencies": {
+    "@ai-sdk/anthropic": "^2.0.17",
+    "@ai-sdk/azure": "^2.0.30",
+    "@ai-sdk/deepseek": "^1.0.17",
+    "@ai-sdk/google": "^2.0.14",
+    "@ai-sdk/openai": "^2.0.30",
+    "@ai-sdk/openai-compatible": "^1.0.17",
+    "@ai-sdk/provider": "^2.0.0",
+    "@ai-sdk/provider-utils": "^3.0.9",
+    "@ai-sdk/xai": "^2.0.18",
+    "zod": "^4.1.5"
+  },
+  "devDependencies": {
+    "tsdown": "^0.12.9",
+    "typescript": "^5.0.0",
+    "vitest": "^3.2.4"
+  },
+  "sideEffects": false,
+  "engines": {
+    "node": ">=18.0.0"
+  },
+  "files": [
+    "dist"
+  ],
+  "exports": {
+    ".": {
+      "types": "./dist/index.d.ts",
+      "react-native": "./dist/index.js",
+      "import": "./dist/index.mjs",
+      "require": "./dist/index.js",
+      "default": "./dist/index.js"
+    },
+    "./built-in/plugins": {
+      "types": "./dist/built-in/plugins/index.d.ts",
+      "react-native": "./dist/built-in/plugins/index.js",
+      "import": "./dist/built-in/plugins/index.mjs",
+      "require": "./dist/built-in/plugins/index.js",
+      "default": "./dist/built-in/plugins/index.js"
+    },
+    "./provider": {
+      "types": "./dist/provider/index.d.ts",
+      "react-native": "./dist/provider/index.js",
+      "import": "./dist/provider/index.mjs",
+      "require": "./dist/provider/index.js",
+      "default": "./dist/provider/index.js"
+    }
+  }
+}

packages/aiCore/setupVitest.ts

@@ -0,0 +1,2 @@
+// 模拟 Vite SSR helper，避免 Node 环境找不到时报错
+;(globalThis as any).__vite_ssr_exportName__ = (name: string, value: any) => value

packages/aiCore/src/core/README.MD

@@ -0,0 +1,3 @@
+# @cherryStudio-aiCore
+
+Core

packages/aiCore/src/core/index.ts

@@ -0,0 +1,17 @@
+/**
+ * Core 模块导出
+ * 内部核心功能，供其他模块使用，不直接面向最终调用者
+ */
+
+// 中间件系统
+export type { NamedMiddleware } from './middleware'
+export { createMiddlewares, wrapModelWithMiddlewares } from './middleware'
+
+// 创建管理
+export { globalModelResolver, ModelResolver } from './models'
+export type { ModelConfig as ModelConfigType } from './models/types'
+
+// 执行管理
+export type { ToolUseRequestContext } from './plugins/built-in/toolUsePlugin/type'
+export { createExecutor, createOpenAICompatibleExecutor } from './runtime'
+export type { RuntimeConfig } from './runtime/types'

packages/aiCore/src/core/middleware/index.ts

@@ -0,0 +1,8 @@
+/**
+ * Middleware 模块导出
+ * 提供通用的中间件管理能力
+ */
+
+export { createMiddlewares } from './manager'
+export type { NamedMiddleware } from './types'
+export { wrapModelWithMiddlewares } from './wrapper'

packages/aiCore/src/core/middleware/manager.ts

@@ -0,0 +1,16 @@
+/**
+ * 中间件管理器
+ * 专注于 AI SDK 中间件的管理，与插件系统分离
+ */
+import { LanguageModelV2Middleware } from '@ai-sdk/provider'
+
+/**
+ * 创建中间件列表
+ * 合并用户提供的中间件
+ */
+export function createMiddlewares(userMiddlewares: LanguageModelV2Middleware[] = []): LanguageModelV2Middleware[] {
+  // 未来可以在这里添加默认的中间件
+  const defaultMiddlewares: LanguageModelV2Middleware[] = []
+
+  return [...defaultMiddlewares, ...userMiddlewares]
+}

packages/aiCore/src/core/middleware/types.ts

@@ -0,0 +1,12 @@
+/**
+ * 中间件系统类型定义
+ */
+import { LanguageModelV2Middleware } from '@ai-sdk/provider'
+
+/**
+ * 具名中间件接口
+ */
+export interface NamedMiddleware {
+  name: string
+  middleware: LanguageModelV2Middleware
+}

packages/aiCore/src/core/middleware/wrapper.ts

@@ -0,0 +1,23 @@
+/**
+ * 模型包装工具函数
+ * 用于将中间件应用到LanguageModel上
+ */
+import { LanguageModelV2, LanguageModelV2Middleware } from '@ai-sdk/provider'
+import { wrapLanguageModel } from 'ai'
+
+/**
+ * 使用中间件包装模型
+ */
+export function wrapModelWithMiddlewares(
+  model: LanguageModelV2,
+  middlewares: LanguageModelV2Middleware[]
+): LanguageModelV2 {
+  if (middlewares.length === 0) {
+    return model
+  }
+
+  return wrapLanguageModel({
+    model,
+    middleware: middlewares
+  })
+}

packages/aiCore/src/core/models/ModelResolver.ts

@@ -0,0 +1,124 @@
+/**
+ * 模型解析器 - models模块的核心
+ * 负责将modelId解析为AI SDK的LanguageModel实例
+ * 支持传统格式和命名空间格式
+ * 集成了来自 ModelCreator 的特殊处理逻辑
+ */
+
+import { EmbeddingModelV2, ImageModelV2, LanguageModelV2, LanguageModelV2Middleware } from '@ai-sdk/provider'
+
+import { wrapModelWithMiddlewares } from '../middleware/wrapper'
+import { DEFAULT_SEPARATOR, globalRegistryManagement } from '../providers/RegistryManagement'
+
+export class ModelResolver {
+  /**
+   * 核心方法：解析任意格式的modelId为语言模型
+   *
+   * @param modelId 模型ID，支持 'gpt-4' 和 'anthropic>claude-3' 两种格式
+   * @param fallbackProviderId 当modelId为传统格式时使用的providerId
+   * @param providerOptions provider配置选项（用于OpenAI模式选择等）
+   * @param middlewares 中间件数组，会应用到最终模型上
+   */
+  async resolveLanguageModel(
+    modelId: string,
+    fallbackProviderId: string,
+    providerOptions?: any,
+    middlewares?: LanguageModelV2Middleware[]
+  ): Promise<LanguageModelV2> {
+    let finalProviderId = fallbackProviderId
+    let model: LanguageModelV2
+    // 🎯 处理 OpenAI 模式选择逻辑 (从 ModelCreator 迁移)
+    if ((fallbackProviderId === 'openai' || fallbackProviderId === 'azure') && providerOptions?.mode === 'chat') {
+      finalProviderId = `${fallbackProviderId}-chat`
+    }
+
+    // 检查是否是命名空间格式
+    if (modelId.includes(DEFAULT_SEPARATOR)) {
+      model = this.resolveNamespacedModel(modelId)
+    } else {
+      // 传统格式：使用处理后的 providerId + modelId
+      model = this.resolveTraditionalModel(finalProviderId, modelId)
+    }
+
+    // 🎯 应用中间件（如果有）
+    if (middlewares && middlewares.length > 0) {
+      model = wrapModelWithMiddlewares(model, middlewares)
+    }
+
+    return model
+  }
+
+  /**
+   * 解析文本嵌入模型
+   */
+  async resolveTextEmbeddingModel(modelId: string, fallbackProviderId: string): Promise<EmbeddingModelV2<string>> {
+    if (modelId.includes(DEFAULT_SEPARATOR)) {
+      return this.resolveNamespacedEmbeddingModel(modelId)
+    }
+
+    return this.resolveTraditionalEmbeddingModel(fallbackProviderId, modelId)
+  }
+
+  /**
+   * 解析图像模型
+   */
+  async resolveImageModel(modelId: string, fallbackProviderId: string): Promise<ImageModelV2> {
+    if (modelId.includes(DEFAULT_SEPARATOR)) {
+      return this.resolveNamespacedImageModel(modelId)
+    }
+
+    return this.resolveTraditionalImageModel(fallbackProviderId, modelId)
+  }
+
+  /**
+   * 解析命名空间格式的语言模型
+   * aihubmix:anthropic:claude-3 -> globalRegistryManagement.languageModel('aihubmix:anthropic:claude-3')
+   */
+  private resolveNamespacedModel(modelId: string): LanguageModelV2 {
+    return globalRegistryManagement.languageModel(modelId as any)
+  }
+
+  /**
+   * 解析传统格式的语言模型
+   * providerId: 'openai', modelId: 'gpt-4' -> globalRegistryManagement.languageModel('openai:gpt-4')
+   */
+  private resolveTraditionalModel(providerId: string, modelId: string): LanguageModelV2 {
+    const fullModelId = `${providerId}${DEFAULT_SEPARATOR}${modelId}`
+    return globalRegistryManagement.languageModel(fullModelId as any)
+  }
+
+  /**
+   * 解析命名空间格式的嵌入模型
+   */
+  private resolveNamespacedEmbeddingModel(modelId: string): EmbeddingModelV2<string> {
+    return globalRegistryManagement.textEmbeddingModel(modelId as any)
+  }
+
+  /**
+   * 解析传统格式的嵌入模型
+   */
+  private resolveTraditionalEmbeddingModel(providerId: string, modelId: string): EmbeddingModelV2<string> {
+    const fullModelId = `${providerId}${DEFAULT_SEPARATOR}${modelId}`
+    return globalRegistryManagement.textEmbeddingModel(fullModelId as any)
+  }
+
+  /**
+   * 解析命名空间格式的图像模型
+   */
+  private resolveNamespacedImageModel(modelId: string): ImageModelV2 {
+    return globalRegistryManagement.imageModel(modelId as any)
+  }
+
+  /**
+   * 解析传统格式的图像模型
+   */
+  private resolveTraditionalImageModel(providerId: string, modelId: string): ImageModelV2 {
+    const fullModelId = `${providerId}${DEFAULT_SEPARATOR}${modelId}`
+    return globalRegistryManagement.imageModel(fullModelId as any)
+  }
+}
+
+/**
+ * 全局模型解析器实例
+ */
+export const globalModelResolver = new ModelResolver()

packages/aiCore/src/core/models/index.ts

@@ -0,0 +1,9 @@
+/**
+ * Models 模块统一导出 - 简化版
+ */
+
+// 核心模型解析器
+export { globalModelResolver, ModelResolver } from './ModelResolver'
+
+// 保留的类型定义（可能被其他地方使用）
+export type { ModelConfig as ModelConfigType } from './types'

packages/aiCore/src/core/models/types.ts

@@ -0,0 +1,15 @@
+/**
+ * Creation 模块类型定义
+ */
+import { LanguageModelV2Middleware } from '@ai-sdk/provider'
+
+import type { ProviderId, ProviderSettingsMap } from '../providers/types'
+
+export interface ModelConfig<T extends ProviderId = ProviderId> {
+  providerId: T
+  modelId: string
+  providerSettings: ProviderSettingsMap[T] & { mode?: 'chat' | 'responses' }
+  middlewares?: LanguageModelV2Middleware[]
+  // 额外模型参数
+  extraModelConfig?: Record<string, any>
+}

packages/aiCore/src/core/options/examples.ts

@@ -0,0 +1,87 @@
+import { streamText } from 'ai'
+
+import {
+  createAnthropicOptions,
+  createGenericProviderOptions,
+  createGoogleOptions,
+  createOpenAIOptions,
+  mergeProviderOptions
+} from './factory'
+
+// 示例1: 使用已知供应商的严格类型约束
+export function exampleOpenAIWithOptions() {
+  const openaiOptions = createOpenAIOptions({
+    reasoningEffort: 'medium'
+  })
+
+  // 这里会有类型检查，确保选项符合OpenAI的设置
+  return streamText({
+    model: {} as any, // 实际使用时替换为真实模型
+    prompt: 'Hello',
+    providerOptions: openaiOptions
+  })
+}
+
+// 示例2: 使用Anthropic供应商选项
+export function exampleAnthropicWithOptions() {
+  const anthropicOptions = createAnthropicOptions({
+    thinking: {
+      type: 'enabled',
+      budgetTokens: 1000
+    }
+  })
+
+  return streamText({
+    model: {} as any,
+    prompt: 'Hello',
+    providerOptions: anthropicOptions
+  })
+}
+
+// 示例3: 使用Google供应商选项
+export function exampleGoogleWithOptions() {
+  const googleOptions = createGoogleOptions({
+    thinkingConfig: {
+      includeThoughts: true,
+      thinkingBudget: 1000
+    }
+  })
+
+  return streamText({
+    model: {} as any,
+    prompt: 'Hello',
+    providerOptions: googleOptions
+  })
+}
+
+// 示例4: 使用未知供应商（通用类型）
+export function exampleUnknownProviderWithOptions() {
+  const customProviderOptions = createGenericProviderOptions('custom-provider', {
+    temperature: 0.7,
+    customSetting: 'value',
+    anotherOption: true
+  })
+
+  return streamText({
+    model: {} as any,
+    prompt: 'Hello',
+    providerOptions: customProviderOptions
+  })
+}
+
+// 示例5: 合并多个供应商选项
+export function exampleMergedOptions() {
+  const openaiOptions = createOpenAIOptions({})
+
+  const customOptions = createGenericProviderOptions('custom', {
+    customParam: 'value'
+  })
+
+  const mergedOptions = mergeProviderOptions(openaiOptions, customOptions)
+
+  return streamText({
+    model: {} as any,
+    prompt: 'Hello',
+    providerOptions: mergedOptions
+  })
+}

packages/aiCore/src/core/options/factory.ts

@@ -0,0 +1,71 @@
+import { ExtractProviderOptions, ProviderOptionsMap, TypedProviderOptions } from './types'
+
+/**
+ * 创建特定供应商的选项
+ * @param provider 供应商名称
+ * @param options 供应商特定的选项
+ * @returns 格式化的provider options
+ */
+export function createProviderOptions<T extends keyof ProviderOptionsMap>(
+  provider: T,
+  options: ExtractProviderOptions<T>
+): Record<T, ExtractProviderOptions<T>> {
+  return { [provider]: options } as Record<T, ExtractProviderOptions<T>>
+}
+
+/**
+ * 创建任意供应商的选项（包括未知供应商）
+ * @param provider 供应商名称
+ * @param options 供应商选项
+ * @returns 格式化的provider options
+ */
+export function createGenericProviderOptions<T extends string>(
+  provider: T,
+  options: Record<string, any>
+): Record<T, Record<string, any>> {
+  return { [provider]: options } as Record<T, Record<string, any>>
+}
+
+/**
+ * 合并多个供应商的options
+ * @param optionsMap 包含多个供应商选项的对象
+ * @returns 合并后的TypedProviderOptions
+ */
+export function mergeProviderOptions(...optionsMap: Partial<TypedProviderOptions>[]): TypedProviderOptions {
+  return Object.assign({}, ...optionsMap)
+}
+
+/**
+ * 创建OpenAI供应商选项的便捷函数
+ */
+export function createOpenAIOptions(options: ExtractProviderOptions<'openai'>) {
+  return createProviderOptions('openai', options)
+}
+
+/**
+ * 创建Anthropic供应商选项的便捷函数
+ */
+export function createAnthropicOptions(options: ExtractProviderOptions<'anthropic'>) {
+  return createProviderOptions('anthropic', options)
+}
+
+/**
+ * 创建Google供应商选项的便捷函数
+ */
+export function createGoogleOptions(options: ExtractProviderOptions<'google'>) {
+  return createProviderOptions('google', options)
+}
+
+/**
+ * 创建OpenRouter供应商选项的便捷函数
+ */
+export function createOpenRouterOptions(options: ExtractProviderOptions<'openrouter'> | Record<string, any>) {
+  return createProviderOptions('openrouter', options)
+}
+
+/**
+ * 创建XAI供应商选项的便捷函数
+ */
+export function createXaiOptions(options: ExtractProviderOptions<'xai'>) {
+  return createProviderOptions('xai', options)
+}

packages/aiCore/src/core/options/index.ts

@@ -0,0 +1,2 @@
+export * from './factory'
+export * from './types'

packages/aiCore/src/core/options/types.ts

@@ -0,0 +1,32 @@
+import { type AnthropicProviderOptions } from '@ai-sdk/anthropic'
+import { type GoogleGenerativeAIProviderOptions } from '@ai-sdk/google'
+import { type OpenAIResponsesProviderOptions } from '@ai-sdk/openai'
+import { type SharedV2ProviderMetadata } from '@ai-sdk/provider'
+import { type XaiProviderOptions } from '@ai-sdk/xai'
+import { type OpenRouterProviderOptions } from '@openrouter/ai-sdk-provider'
+
+export type ProviderOptions<T extends keyof SharedV2ProviderMetadata> = SharedV2ProviderMetadata[T]
+
+/**
+ * 供应商选项类型，如果map中没有，说明没有约束
+ */
+export type ProviderOptionsMap = {
+  openai: OpenAIResponsesProviderOptions
+  anthropic: AnthropicProviderOptions
+  google: GoogleGenerativeAIProviderOptions
+  openrouter: OpenRouterProviderOptions
+  xai: XaiProviderOptions
+}
+
+// 工具类型，用于从ProviderOptionsMap中提取特定供应商的选项类型
+export type ExtractProviderOptions<T extends keyof ProviderOptionsMap> = ProviderOptionsMap[T]
+
+/**
+ * 类型安全的ProviderOptions
+ * 对于已知供应商使用严格类型，对于未知供应商允许任意Record<string, JSONValue>
+ */
+export type TypedProviderOptions = {
+  [K in keyof ProviderOptionsMap]?: ProviderOptionsMap[K]
+} & {
+  [K in string]?: Record<string, any>
+} & SharedV2ProviderMetadata

packages/aiCore/src/core/plugins/README.md

@@ -0,0 +1,257 @@
+# AI Core 插件系统
+
+支持四种钩子类型：**First**、**Sequential**、**Parallel** 和 **Stream**。
+
+## 🎯 设计理念
+
+- **语义清晰**：不同钩子有不同的执行语义
+- **类型安全**：TypeScript 完整支持
+- **性能优化**：First 短路、Parallel 并发、Sequential 链式
+- **易于扩展**：`enforce` 排序 + 功能分类
+
+## 📋 钩子类型
+
+### 🥇 First 钩子 - 首个有效结果
+
+```typescript
+// 只执行第一个返回值的插件，用于解析和查找
+resolveModel?: (modelId: string, context: AiRequestContext) => string | null
+loadTemplate?: (templateName: string, context: AiRequestContext) => any | null
+```
+
+### 🔄 Sequential 钩子 - 链式数据转换
+
+```typescript
+// 按顺序链式执行，每个插件可以修改数据
+transformParams?: (params: any, context: AiRequestContext) => any
+transformResult?: (result: any, context: AiRequestContext) => any
+```
+
+### ⚡ Parallel 钩子 - 并行副作用
+
+```typescript
+// 并发执行，用于日志、监控等副作用
+onRequestStart?: (context: AiRequestContext) => void
+onRequestEnd?: (context: AiRequestContext, result: any) => void
+onError?: (error: Error, context: AiRequestContext) => void
+```
+
+### 🌊 Stream 钩子 - 流处理
+
+```typescript
+// 直接使用 AI SDK 的 TransformStream
+transformStream?: () => (options) => TransformStream<TextStreamPart, TextStreamPart>
+```
+
+## 🚀 快速开始
+
+### 基础用法
+
+```typescript
+import { PluginManager, createContext, definePlugin } from '@cherrystudio/ai-core/middleware'
+
+// 创建插件管理器
+const pluginManager = new PluginManager()
+
+// 添加插件
+pluginManager.use({
+  name: 'my-plugin',
+  async transformParams(params, context) {
+    return { ...params, temperature: 0.7 }
+  }
+})
+
+// 使用插件
+const context = createContext('openai', 'gpt-4', { messages: [] })
+const transformedParams = await pluginManager.executeSequential(
+  'transformParams',
+  { messages: [{ role: 'user', content: 'Hello' }] },
+  context
+)
+```
+
+### 完整示例
+
+```typescript
+import {
+  PluginManager,
+  ModelAliasPlugin,
+  LoggingPlugin,
+  ParamsValidationPlugin,
+  createContext
+} from '@cherrystudio/ai-core/middleware'
+
+// 创建插件管理器
+const manager = new PluginManager([
+  ModelAliasPlugin, // 模型别名解析
+  ParamsValidationPlugin, // 参数验证
+  LoggingPlugin // 日志记录
+])
+
+// AI 请求流程
+async function aiRequest(providerId: string, modelId: string, params: any) {
+  const context = createContext(providerId, modelId, params)
+
+  try {
+    // 1. 【并行】触发请求开始事件
+    await manager.executeParallel('onRequestStart', context)
+
+    // 2. 【首个】解析模型别名
+    const resolvedModel = await manager.executeFirst('resolveModel', modelId, context)
+    context.modelId = resolvedModel || modelId
+
+    // 3. 【串行】转换请求参数
+    const transformedParams = await manager.executeSequential('transformParams', params, context)
+
+    // 4. 【流处理】收集流转换器（AI SDK 原生支持数组）
+    const streamTransforms = manager.collectStreamTransforms()
+
+    // 5. 调用 AI SDK（这里省略具体实现）
+    const result = await callAiSdk(transformedParams, streamTransforms)
+
+    // 6. 【串行】转换响应结果
+    const transformedResult = await manager.executeSequential('transformResult', result, context)
+
+    // 7. 【并行】触发请求完成事件
+    await manager.executeParallel('onRequestEnd', context, transformedResult)
+
+    return transformedResult
+  } catch (error) {
+    // 8. 【并行】触发错误事件
+    await manager.executeParallel('onError', context, undefined, error)
+    throw error
+  }
+}
+```
+
+## 🔧 自定义插件
+
+### 模型别名插件
+
+```typescript
+const ModelAliasPlugin = definePlugin({
+  name: 'model-alias',
+  enforce: 'pre', // 最先执行
+
+  async resolveModel(modelId) {
+    const aliases = {
+      gpt4: 'gpt-4-turbo-preview',
+      claude: 'claude-3-sonnet-20240229'
+    }
+    return aliases[modelId] || null
+  }
+})
+```
+
+### 参数验证插件
+
+```typescript
+const ValidationPlugin = definePlugin({
+  name: 'validation',
+
+  async transformParams(params) {
+    if (!params.messages) {
+      throw new Error('messages is required')
+    }
+
+    return {
+      ...params,
+      temperature: params.temperature ?? 0.7,
+      max_tokens: params.max_tokens ?? 4096
+    }
+  }
+})
+```
+
+### 监控插件
+
+```typescript
+const MonitoringPlugin = definePlugin({
+  name: 'monitoring',
+  enforce: 'post', // 最后执行
+
+  async onRequestEnd(context, result) {
+    const duration = Date.now() - context.startTime
+    console.log(`请求耗时: ${duration}ms`)
+  }
+})
+```
+
+### 内容过滤插件
+
+```typescript
+const FilterPlugin = definePlugin({
+  name: 'content-filter',
+
+  transformStream() {
+    return () =>
+      new TransformStream({
+        transform(chunk, controller) {
+          if (chunk.type === 'text-delta') {
+            const filtered = chunk.textDelta.replace(/敏感词/g, '***')
+            controller.enqueue({ ...chunk, textDelta: filtered })
+          } else {
+            controller.enqueue(chunk)
+          }
+        }
+      })
+  }
+})
+```
+
+## 📊 执行顺序
+
+### 插件排序
+
+```
+enforce: 'pre' → normal → enforce: 'post'
+```
+
+### 钩子执行流程
+
+```mermaid
+graph TD
+    A[请求开始] --> B[onRequestStart 并行执行]
+    B --> C[resolveModel 首个有效]
+    C --> D[loadTemplate 首个有效]
+    D --> E[transformParams 串行执行]
+    E --> F[collectStreamTransforms]
+    F --> G[AI SDK 调用]
+    G --> H[transformResult 串行执行]
+    H --> I[onRequestEnd 并行执行]
+
+    G --> J[异常处理]
+    J --> K[onError 并行执行]
+```
+
+## 💡 最佳实践
+
+1. **功能单一**：每个插件专注一个功能
+2. **幂等性**：插件应该是幂等的，重复执行不会产生副作用
+3. **错误处理**：插件内部处理异常，不要让异常向上传播
+4. **性能优化**：使用合适的钩子类型（First vs Sequential vs Parallel）
+5. **命名规范**：使用语义化的插件名称
+
+## 🔍 调试工具
+
+```typescript
+// 查看插件统计信息
+const stats = manager.getStats()
+console.log('插件统计:', stats)
+
+// 查看所有插件
+const plugins = manager.getPlugins()
+console.log(
+  '已注册插件:',
+  plugins.map((p) => p.name)
+)
+```
+
+## ⚡ 性能优势
+
+- **First 钩子**：一旦找到结果立即停止，避免无效计算
+- **Parallel 钩子**：真正并发执行，不阻塞主流程
+- **Sequential 钩子**：保证数据转换的顺序性
+- **Stream 钩子**：直接集成 AI SDK，零开销
+
+这个设计兼顾了简洁性和强大功能，为 AI Core 提供了灵活而高效的扩展机制。

packages/aiCore/src/core/plugins/built-in/googleToolsPlugin/index.ts

@@ -0,0 +1,38 @@
+import { google } from '@ai-sdk/google'
+
+import { definePlugin } from '../../'
+import type { AiRequestContext } from '../../types'
+
+const toolNameMap = {
+  googleSearch: 'google_search',
+  urlContext: 'url_context',
+  codeExecution: 'code_execution'
+} as const
+
+type ToolConfigKey = keyof typeof toolNameMap
+type ToolConfig = { googleSearch?: boolean; urlContext?: boolean; codeExecution?: boolean }
+
+export const googleToolsPlugin = (config?: ToolConfig) =>
+  definePlugin({
+    name: 'googleToolsPlugin',
+    transformParams: <T>(params: T, context: AiRequestContext): T => {
+      const { providerId } = context
+      if (providerId === 'google' && config) {
+        if (typeof params === 'object' && params !== null) {
+          const typedParams = params as T & { tools?: Record<string, unknown> }
+
+          if (!typedParams.tools) {
+            typedParams.tools = {}
+          }
+          // 使用类型安全的方式遍历配置
+          ;(Object.keys(config) as ToolConfigKey[]).forEach((key) => {
+            if (config[key] && key in toolNameMap && key in google.tools) {
+              const toolName = toolNameMap[key]
+              typedParams.tools![toolName] = google.tools[key]({})
+            }
+          })
+        }
+      }
+      return params
+    }
+  })

packages/aiCore/src/core/plugins/built-in/index.ts

@@ -0,0 +1,15 @@
+/**
+ * 内置插件命名空间
+ * 所有内置插件都以 'built-in:' 为前缀
+ */
+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'

packages/aiCore/src/core/plugins/built-in/logging.ts

@@ -0,0 +1,86 @@
+/**
+ * 内置插件：日志记录
+ * 记录AI调用的关键信息，支持性能监控和调试
+ */
+import { definePlugin } from '../index'
+import type { AiRequestContext } from '../types'
+
+export interface LoggingConfig {
+  // 日志级别
+  level?: 'debug' | 'info' | 'warn' | 'error'
+  // 是否记录参数
+  logParams?: boolean
+  // 是否记录结果
+  logResult?: boolean
+  // 是否记录性能数据
+  logPerformance?: boolean
+  // 自定义日志函数
+  logger?: (level: string, message: string, data?: any) => void
+}
+
+/**
+ * 创建日志插件
+ */
+export function createLoggingPlugin(config: LoggingConfig = {}) {
+  const { level = 'info', logParams = true, logResult = false, logPerformance = true, logger = console.log } = config
+
+  const startTimes = new Map<string, number>()
+
+  return definePlugin({
+    name: 'built-in:logging',
+
+    onRequestStart: (context: AiRequestContext) => {
+      const requestId = context.requestId
+      startTimes.set(requestId, Date.now())
+
+      logger(level, `🚀 AI Request Started`, {
+        requestId,
+        providerId: context.providerId,
+        modelId: context.modelId,
+        originalParams: logParams ? context.originalParams : '[hidden]'
+      })
+    },
+
+    onRequestEnd: (context: AiRequestContext, result: any) => {
+      const requestId = context.requestId
+      const startTime = startTimes.get(requestId)
+      const duration = startTime ? Date.now() - startTime : undefined
+      startTimes.delete(requestId)
+
+      const logData: any = {
+        requestId,
+        providerId: context.providerId,
+        modelId: context.modelId
+      }
+
+      if (logPerformance && duration) {
+        logData.duration = `${duration}ms`
+      }
+
+      if (logResult) {
+        logData.result = result
+      }
+
+      logger(level, `✅ AI Request Completed`, logData)
+    },
+
+    onError: (error: Error, context: AiRequestContext) => {
+      const requestId = context.requestId
+      const startTime = startTimes.get(requestId)
+      const duration = startTime ? Date.now() - startTime : undefined
+      startTimes.delete(requestId)
+
+      logger('error', `❌ AI Request Failed`, {
+        requestId,
+        providerId: context.providerId,
+        modelId: context.modelId,
+        duration: duration ? `${duration}ms` : undefined,
+        error: {
+          name: error.name,
+          message: error.message,
+          stack: error.stack
+        }
+      })
+    }
+  })
+}

packages/aiCore/src/core/plugins/built-in/toolUsePlugin/StreamEventManager.ts

@@ -0,0 +1,174 @@
+/**
+ * 流事件管理器
+ *
+ * 负责处理 AI SDK 流事件的发送和管理
+ * 从 promptToolUsePlugin.ts 中提取出来以降低复杂度
+ */
+import type { ModelMessage } from 'ai'
+
+import type { AiRequestContext } from '../../types'
+import type { StreamController } from './ToolExecutor'
+
+/**
+ * 流事件管理器类
+ */
+export class StreamEventManager {
+  /**
+   * 发送工具调用步骤开始事件
+   */
+  sendStepStartEvent(controller: StreamController): void {
+    controller.enqueue({
+      type: 'start-step',
+      request: {},
+      warnings: []
+    })
+  }
+
+  /**
+   * 发送步骤完成事件
+   */
+  sendStepFinishEvent(
+    controller: StreamController,
+    chunk: any,
+    context: AiRequestContext,
+    finishReason: string = 'stop'
+  ): void {
+    // 累加当前步骤的 usage
+    if (chunk.usage && context.accumulatedUsage) {
+      this.accumulateUsage(context.accumulatedUsage, chunk.usage)
+    }
+
+    controller.enqueue({
+      type: 'finish-step',
+      finishReason,
+      response: chunk.response,
+      usage: chunk.usage,
+      providerMetadata: chunk.providerMetadata
+    })
+  }
+
+  /**
+   * 处理递归调用并将结果流接入当前流
+   */
+  async handleRecursiveCall(
+    controller: StreamController,
+    recursiveParams: any,
+    context: AiRequestContext
+  ): Promise<void> {
+    // try {
+    // 重置工具执行状态，准备处理新的步骤
+    context.hasExecutedToolsInCurrentStep = false
+
+    const recursiveResult = await context.recursiveCall(recursiveParams)
+
+    if (recursiveResult && recursiveResult.fullStream) {
+      await this.pipeRecursiveStream(controller, recursiveResult.fullStream, context)
+    } else {
+      console.warn('[MCP Prompt] No fullstream found in recursive result:', recursiveResult)
+    }
+    // } catch (error) {
+    //   this.handleRecursiveCallError(controller, error, stepId)
+    // }
+  }
+
+  /**
+   * 将递归流的数据传递到当前流
+   */
+  private async pipeRecursiveStream(
+    controller: StreamController,
+    recursiveStream: ReadableStream,
+    context?: AiRequestContext
+  ): Promise<void> {
+    const reader = recursiveStream.getReader()
+    try {
+      while (true) {
+        const { done, value } = await reader.read()
+        if (done) {
+          break
+        }
+        if (value.type === 'finish') {
+          // 迭代的流不发finish，但需要累加其 usage
+          if (value.usage && context?.accumulatedUsage) {
+            this.accumulateUsage(context.accumulatedUsage, value.usage)
+          }
+          break
+        }
+        // 对于 finish-step 类型，累加其 usage
+        if (value.type === 'finish-step' && value.usage && context?.accumulatedUsage) {
+          this.accumulateUsage(context.accumulatedUsage, value.usage)
+        }
+        // 将递归流的数据传递到当前流
+        controller.enqueue(value)
+      }
+    } finally {
+      reader.releaseLock()
+    }
+  }
+
+  /**
+   * 处理递归调用错误
+   */
+  // private handleRecursiveCallError(controller: StreamController, error: unknown): void {
+  //   console.error('[MCP Prompt] Recursive call failed:', error)
+
+  //   // 使用 AI SDK 标准错误格式，但不中断流
+  //   controller.enqueue({
+  //     type: 'error',
+  //     error: {
+  //       message: error instanceof Error ? error.message : String(error),
+  //       name: error instanceof Error ? error.name : 'RecursiveCallError'
+  //     }
+  //   })
+
+  //   // // 继续发送文本增量，保持流的连续性
+  //   // controller.enqueue({
+  //   //   type: 'text-delta',
+  //   //   id: stepId,
+  //   //   text: '\n\n[工具执行后递归调用失败，继续对话...]'
+  //   // })
+  // }
+
+  /**
+   * 构建递归调用的参数
+   */
+  buildRecursiveParams(context: AiRequestContext, textBuffer: string, toolResultsText: string, tools: any): any {
+    // 构建新的对话消息
+    const newMessages: ModelMessage[] = [
+      ...(context.originalParams.messages || []),
+      {
+        role: 'assistant',
+        content: textBuffer
+      },
+      {
+        role: 'user',
+        content: toolResultsText
+      }
+    ]
+
+    // 递归调用，继续对话，重新传递 tools
+    const recursiveParams = {
+      ...context.originalParams,
+      messages: newMessages,
+      tools: tools
+    }
+
+    // 更新上下文中的消息
+    context.originalParams.messages = newMessages
+
+    return recursiveParams
+  }
+
+  /**
+   * 累加 usage 数据
+   */
+  private accumulateUsage(target: any, source: any): void {
+    if (!target || !source) return
+
+    // 累加各种 token 类型
+    target.inputTokens = (target.inputTokens || 0) + (source.inputTokens || 0)
+    target.outputTokens = (target.outputTokens || 0) + (source.outputTokens || 0)
+    target.totalTokens = (target.totalTokens || 0) + (source.totalTokens || 0)
+    target.reasoningTokens = (target.reasoningTokens || 0) + (source.reasoningTokens || 0)
+    target.cachedInputTokens = (target.cachedInputTokens || 0) + (source.cachedInputTokens || 0)
+  }
+}

packages/aiCore/src/core/plugins/built-in/toolUsePlugin/ToolExecutor.ts

@@ -0,0 +1,151 @@
+/**
+ * 工具执行器
+ *
+ * 负责工具的执行、结果格式化和相关事件发送
+ * 从 promptToolUsePlugin.ts 中提取出来以降低复杂度
+ */
+import type { ToolSet, TypedToolError } from 'ai'
+
+import type { ToolUseResult } from './type'
+
+/**
+ * 工具执行结果
+ */
+export interface ExecutedResult {
+  toolCallId: string
+  toolName: string
+  result: any
+  isError?: boolean
+}
+
+/**
+ * 流控制器类型（从 AI SDK 提取）
+ */
+export interface StreamController {
+  enqueue(chunk: any): void
+}
+
+/**
+ * 工具执行器类
+ */
+export class ToolExecutor {
+  /**
+   * 执行多个工具调用
+   */
+  async executeTools(
+    toolUses: ToolUseResult[],
+    tools: ToolSet,
+    controller: StreamController
+  ): Promise<ExecutedResult[]> {
+    const executedResults: ExecutedResult[] = []
+    for (const toolUse of toolUses) {
+      try {
+        const tool = tools[toolUse.toolName]
+        if (!tool || typeof tool.execute !== 'function') {
+          throw new Error(`Tool "${toolUse.toolName}" has no execute method`)
+        }
+
+        // 发送 tool-call 事件
+        controller.enqueue({
+          type: 'tool-call',
+          toolCallId: toolUse.id,
+          toolName: toolUse.toolName,
+          input: toolUse.arguments
+        })
+
+        const result = await tool.execute(toolUse.arguments, {
+          toolCallId: toolUse.id,
+          messages: [],
+          abortSignal: new AbortController().signal
+        })
+
+        // 发送 tool-result 事件
+        controller.enqueue({
+          type: 'tool-result',
+          toolCallId: toolUse.id,
+          toolName: toolUse.toolName,
+          input: toolUse.arguments,
+          output: result
+        })
+
+        executedResults.push({
+          toolCallId: toolUse.id,
+          toolName: toolUse.toolName,
+          result,
+          isError: false
+        })
+      } catch (error) {
+        console.error(`[MCP Prompt Stream] Tool execution failed: ${toolUse.toolName}`, error)
+
+        // 处理错误情况
+        const errorResult = this.handleToolError(toolUse, error, controller)
+        executedResults.push(errorResult)
+      }
+    }
+
+    return executedResults
+  }
+
+  /**
+   * 格式化工具结果为 Cherry Studio 标准格式
+   */
+  formatToolResults(executedResults: ExecutedResult[]): string {
+    return executedResults
+      .map((tr) => {
+        if (!tr.isError) {
+          return `<tool_use_result>\n  <name>${tr.toolName}</name>\n  <result>${JSON.stringify(tr.result)}</result>\n</tool_use_result>`
+        } else {
+          const error = tr.result || 'Unknown error'
+          return `<tool_use_result>\n  <name>${tr.toolName}</name>\n  <error>${error}</error>\n</tool_use_result>`
+        }
+      })
+      .join('\n\n')
+  }
+
+  /**
+   * 发送工具调用开始相关事件
+   */
+  // private sendToolStartEvents(controller: StreamController, toolUse: ToolUseResult): void {
+  //   // 发送 tool-input-start 事件
+  //   controller.enqueue({
+  //     type: 'tool-input-start',
+  //     id: toolUse.id,
+  //     toolName: toolUse.toolName
+  //   })
+  // }
+
+  /**
+   * 处理工具执行错误
+   */
+  private handleToolError<T extends ToolSet>(
+    toolUse: ToolUseResult,
+    error: unknown,
+    controller: StreamController
+  ): ExecutedResult {
+    // 使用 AI SDK 标准错误格式
+    const toolError: TypedToolError<T> = {
+      type: 'tool-error',
+      toolCallId: toolUse.id,
+      toolName: toolUse.toolName,
+      input: toolUse.arguments,
+      error
+    }
+
+    controller.enqueue(toolError)
+
+    // 发送标准错误事件
+    // controller.enqueue({
+    //   type: 'tool-error',
+    //   toolCallId: toolUse.id,
+    //   error: error instanceof Error ? error.message : String(error),
+    //   input: toolUse.arguments
+    // })
+
+    return {
+      toolCallId: toolUse.id,
+      toolName: toolUse.toolName,
+      result: error,
+      isError: true
+    }
+  }
+}

packages/aiCore/src/core/plugins/built-in/toolUsePlugin/promptToolUsePlugin.ts

@@ -0,0 +1,427 @@
+/**
+ * 内置插件：MCP Prompt 模式
+ * 为不支持原生 Function Call 的模型提供 prompt 方式的工具调用
+ * 内置默认逻辑，支持自定义覆盖
+ */
+import type { TextStreamPart, ToolSet } from 'ai'
+
+import { definePlugin } from '../../index'
+import type { AiRequestContext } from '../../types'
+import { StreamEventManager } from './StreamEventManager'
+import { type TagConfig, TagExtractor } from './tagExtraction'
+import { ToolExecutor } from './ToolExecutor'
+import { PromptToolUseConfig, ToolUseResult } from './type'
+
+/**
+ * 工具使用标签配置
+ */
+const TOOL_USE_TAG_CONFIG: TagConfig = {
+  openingTag: '<tool_use>',
+  closingTag: '</tool_use>',
+  separator: '\n'
+}
+
+/**
+ * 默认系统提示符模板（提取自 Cherry Studio）
+ */
+const DEFAULT_SYSTEM_PROMPT = `In this environment you have access to a set of tools you can use to answer the user's question. \\
+You can use one tool per message, and will receive the result of that tool use in the user's response. You use tools step-by-step to accomplish a given task, with each tool use informed by the result of the previous tool use.
+
+## Tool Use Formatting
+
+Tool use is formatted using XML-style tags. The tool name is enclosed in opening and closing tags, and each parameter is similarly enclosed within its own set of tags. Here's the structure:
+
+<tool_use>
+  <name>{tool_name}</name>
+  <arguments>{json_arguments}</arguments>
+</tool_use>
+
+The tool name should be the exact name of the tool you are using, and the arguments should be a JSON object containing the parameters required by that tool. For example:
+<tool_use>
+  <name>python_interpreter</name>
+  <arguments>{"code": "5 + 3 + 1294.678"}</arguments>
+</tool_use>
+
+The user will respond with the result of the tool use, which should be formatted as follows:
+
+<tool_use_result>
+  <name>{tool_name}</name>
+  <result>{result}</result>
+</tool_use_result>
+
+The result should be a string, which can represent a file or any other output type. You can use this result as input for the next action.
+For example, if the result of the tool use is an image file, you can use it in the next action like this:
+
+<tool_use>
+  <name>image_transformer</name>
+  <arguments>{"image": "image_1.jpg"}</arguments>
+</tool_use>
+
+Always adhere to this format for the tool use to ensure proper parsing and execution.
+
+## Tool Use Examples
+{{ TOOL_USE_EXAMPLES }}
+
+## Tool Use Available Tools
+Above example were using notional tools that might not exist for you. You only have access to these tools:
+{{ AVAILABLE_TOOLS }}
+
+## Tool Use Rules
+Here are the rules you should always follow to solve your task:
+1. Always use the right arguments for the tools. Never use variable names as the action arguments, use the value instead.
+2. Call a tool only when needed: do not call the search agent if you do not need information, try to solve the task yourself.
+3. If no tool call is needed, just answer the question directly.
+4. Never re-do a tool call that you previously did with the exact same parameters.
+5. For tool use, MAKE SURE use XML tag format as shown in the examples above. Do not use any other format.
+
+# User Instructions
+{{ USER_SYSTEM_PROMPT }}
+
+Now Begin! If you solve the task correctly, you will receive a reward of $1,000,000.`
+
+/**
+ * 默认工具使用示例（提取自 Cherry Studio）
+ */
+const DEFAULT_TOOL_USE_EXAMPLES = `
+Here are a few examples using notional tools:
+---
+User: Generate an image of the oldest person in this document.
+
+A: I can use the document_qa tool to find out who the oldest person is in the document.
+<tool_use>
+  <name>document_qa</name>
+  <arguments>{"document": "document.pdf", "question": "Who is the oldest person mentioned?"}</arguments>
+</tool_use>
+
+User: <tool_use_result>
+  <name>document_qa</name>
+  <result>John Doe, a 55 year old lumberjack living in Newfoundland.</result>
+</tool_use_result>
+
+A: I can use the image_generator tool to create a portrait of John Doe.
+<tool_use>
+  <name>image_generator</name>
+  <arguments>{"prompt": "A portrait of John Doe, a 55-year-old man living in Canada."}</arguments>
+</tool_use>
+
+User: <tool_use_result>
+  <name>image_generator</name>
+  <result>image.png</result>
+</tool_use_result>
+
+A: the image is generated as image.png
+
+---
+User: "What is the result of the following operation: 5 + 3 + 1294.678?"
+
+A: I can use the python_interpreter tool to calculate the result of the operation.
+<tool_use>
+  <name>python_interpreter</name>
+  <arguments>{"code": "5 + 3 + 1294.678"}</arguments>
+</tool_use>
+
+User: <tool_use_result>
+  <name>python_interpreter</name>
+  <result>1302.678</result>
+</tool_use_result>
+
+A: The result of the operation is 1302.678.
+
+---
+User: "Which city has the highest population , Guangzhou or Shanghai?"
+
+A: I can use the search tool to find the population of Guangzhou.
+<tool_use>
+  <name>search</name>
+  <arguments>{"query": "Population Guangzhou"}</arguments>
+</tool_use>
+
+User: <tool_use_result>
+  <name>search</name>
+  <result>Guangzhou has a population of 15 million inhabitants as of 2021.</result>
+</tool_use_result>
+
+A: I can use the search tool to find the population of Shanghai.
+<tool_use>
+  <name>search</name>
+  <arguments>{"query": "Population Shanghai"}</arguments>
+</tool_use>
+
+User: <tool_use_result>
+  <name>search</name>
+  <result>26 million (2019)</result>
+</tool_use_result>
+Assistant: The population of Shanghai is 26 million, while Guangzhou has a population of 15 million. Therefore, Shanghai has the highest population.`
+
+/**
+ * 构建可用工具部分（提取自 Cherry Studio）
+ */
+function buildAvailableTools(tools: ToolSet): string | null {
+  const availableTools = Object.keys(tools)
+  if (availableTools.length === 0) return null
+  const result = availableTools
+    .map((toolName: string) => {
+      const tool = tools[toolName]
+      return `
+<tool>
+  <name>${toolName}</name>
+  <description>${tool.description || ''}</description>
+  <arguments>
+    ${tool.inputSchema ? JSON.stringify(tool.inputSchema) : ''}
+  </arguments>
+</tool>
+`
+    })
+    .join('\n')
+  return `<tools>
+${result}
+</tools>`
+}
+
+/**
+ * 默认的系统提示符构建函数（提取自 Cherry Studio）
+ */
+function defaultBuildSystemPrompt(userSystemPrompt: string, tools: ToolSet): string {
+  const availableTools = buildAvailableTools(tools)
+  if (availableTools === null) return userSystemPrompt
+
+  const fullPrompt = DEFAULT_SYSTEM_PROMPT.replace('{{ TOOL_USE_EXAMPLES }}', DEFAULT_TOOL_USE_EXAMPLES)
+    .replace('{{ AVAILABLE_TOOLS }}', availableTools)
+    .replace('{{ USER_SYSTEM_PROMPT }}', userSystemPrompt || '')
+
+  return fullPrompt
+}
+
+/**
+ * 默认工具解析函数（提取自 Cherry Studio）
+ * 解析 XML 格式的工具调用
+ */
+function defaultParseToolUse(content: string, tools: ToolSet): { results: ToolUseResult[]; content: string } {
+  if (!content || !tools || Object.keys(tools).length === 0) {
+    return { results: [], content: content }
+  }
+
+  // 支持两种格式：
+  // 1. 完整的 <tool_use></tool_use> 标签包围的内容
+  // 2. 只有内部内容（从 TagExtractor 提取出来的）
+
+  let contentToProcess = content
+  // 如果内容不包含 <tool_use> 标签，说明是从 TagExtractor 提取的内部内容，需要包装
+  if (!content.includes('<tool_use>')) {
+    contentToProcess = `<tool_use>\n${content}\n</tool_use>`
+  }
+
+  const toolUsePattern =
+    /<tool_use>([\s\S]*?)<name>([\s\S]*?)<\/name>([\s\S]*?)<arguments>([\s\S]*?)<\/arguments>([\s\S]*?)<\/tool_use>/g
+  const results: ToolUseResult[] = []
+  let match
+  let idx = 0
+
+  // Find all tool use blocks
+  while ((match = toolUsePattern.exec(contentToProcess)) !== null) {
+    const fullMatch = match[0]
+    const toolName = match[2].trim()
+    const toolArgs = match[4].trim()
+
+    // Try to parse the arguments as JSON
+    let parsedArgs
+    try {
+      parsedArgs = JSON.parse(toolArgs)
+    } catch (error) {
+      // If parsing fails, use the string as is
+      parsedArgs = toolArgs
+    }
+
+    // Find the corresponding tool
+    const tool = tools[toolName]
+    if (!tool) {
+      console.warn(`Tool "${toolName}" not found in available tools`)
+      continue
+    }
+
+    // Add to results array
+    results.push({
+      id: `${toolName}-${idx++}`, // Unique ID for each tool use
+      toolName: toolName,
+      arguments: parsedArgs,
+      status: 'pending'
+    })
+    contentToProcess = contentToProcess.replace(fullMatch, '')
+  }
+  return { results, content: contentToProcess }
+}
+
+export const createPromptToolUsePlugin = (config: PromptToolUseConfig = {}) => {
+  const { enabled = true, buildSystemPrompt = defaultBuildSystemPrompt, parseToolUse = defaultParseToolUse } = config
+
+  return definePlugin({
+    name: 'built-in:prompt-tool-use',
+    transformParams: (params: any, context: AiRequestContext) => {
+      if (!enabled || !params.tools || typeof params.tools !== 'object') {
+        return params
+      }
+
+      context.mcpTools = params.tools
+
+      // 构建系统提示符
+      const userSystemPrompt = typeof params.system === 'string' ? params.system : ''
+      const systemPrompt = buildSystemPrompt(userSystemPrompt, params.tools)
+      let systemMessage: string | null = systemPrompt
+      if (config.createSystemMessage) {
+        // 🎯 如果用户提供了自定义处理函数，使用它
+        systemMessage = config.createSystemMessage(systemPrompt, params, context)
+      }
+
+      // 移除 tools，改为 prompt 模式
+      const transformedParams = {
+        ...params,
+        ...(systemMessage ? { system: systemMessage } : {}),
+        tools: undefined
+      }
+      context.originalParams = transformedParams
+      return transformedParams
+    },
+    transformStream: (_: any, context: AiRequestContext) => () => {
+      let textBuffer = ''
+      // let stepId = ''
+
+      if (!context.mcpTools) {
+        throw new Error('No tools available')
+      }
+
+      // 从 context 中获取或初始化 usage 累加器
+      if (!context.accumulatedUsage) {
+        context.accumulatedUsage = {
+          inputTokens: 0,
+          outputTokens: 0,
+          totalTokens: 0,
+          reasoningTokens: 0,
+          cachedInputTokens: 0
+        }
+      }
+
+      // 创建工具执行器、流事件管理器和标签提取器
+      const toolExecutor = new ToolExecutor()
+      const streamEventManager = new StreamEventManager()
+      const tagExtractor = new TagExtractor(TOOL_USE_TAG_CONFIG)
+
+      // 在context中初始化工具执行状态，避免递归调用时状态丢失
+      if (!context.hasExecutedToolsInCurrentStep) {
+        context.hasExecutedToolsInCurrentStep = false
+      }
+
+      // 用于hold text-start事件，直到确认有非工具标签内容
+      let pendingTextStart: TextStreamPart<TOOLS> | null = null
+      let hasStartedText = false
+
+      type TOOLS = NonNullable<typeof context.mcpTools>
+      return new TransformStream<TextStreamPart<TOOLS>, TextStreamPart<TOOLS>>({
+        async transform(
+          chunk: TextStreamPart<TOOLS>,
+          controller: TransformStreamDefaultController<TextStreamPart<TOOLS>>
+        ) {
+          // Hold住text-start事件，直到确认有非工具标签内容
+          if ((chunk as any).type === 'text-start') {
+            pendingTextStart = chunk
+            return
+          }
+
+          // text-delta阶段：收集文本内容并过滤工具标签
+          if (chunk.type === 'text-delta') {
+            textBuffer += chunk.text || ''
+            // stepId = chunk.id || ''
+
+            // 使用TagExtractor过滤工具标签，只传递非标签内容到UI层
+            const extractionResults = tagExtractor.processText(chunk.text || '')
+
+            for (const result of extractionResults) {
+              // 只传递非标签内容到UI层
+              if (!result.isTagContent && result.content) {
+                // 如果还没有发送text-start且有pending的text-start，先发送它
+                if (!hasStartedText && pendingTextStart) {
+                  controller.enqueue(pendingTextStart)
+                  hasStartedText = true
+                  pendingTextStart = null
+                }
+
+                const filteredChunk = {
+                  ...chunk,
+                  text: result.content
+                }
+                controller.enqueue(filteredChunk)
+              }
+            }
+            return
+          }
+
+          if (chunk.type === 'text-end') {
+            // 只有当已经发送了text-start时才发送text-end
+            if (hasStartedText) {
+              controller.enqueue(chunk)
+            }
+            return
+          }
+
+          if (chunk.type === 'finish-step') {
+            // 统一在finish-step阶段检查并执行工具调用
+            const tools = context.mcpTools
+            if (tools && Object.keys(tools).length > 0 && !context.hasExecutedToolsInCurrentStep) {
+              // 解析完整的textBuffer来检测工具调用
+              const { results: parsedTools } = parseToolUse(textBuffer, tools)
+              const validToolUses = parsedTools.filter((t) => t.status === 'pending')
+
+              if (validToolUses.length > 0) {
+                context.hasExecutedToolsInCurrentStep = true
+
+                // 执行工具调用（不需要手动发送 start-step，外部流已经处理）
+                const executedResults = await toolExecutor.executeTools(validToolUses, tools, controller)
+
+                // 发送步骤完成事件，使用 tool-calls 作为 finishReason
+                streamEventManager.sendStepFinishEvent(controller, chunk, context, 'tool-calls')
+
+                // 处理递归调用
+                const toolResultsText = toolExecutor.formatToolResults(executedResults)
+                const recursiveParams = streamEventManager.buildRecursiveParams(
+                  context,
+                  textBuffer,
+                  toolResultsText,
+                  tools
+                )
+
+                await streamEventManager.handleRecursiveCall(controller, recursiveParams, context)
+                return
+              }
+            }
+
+            // 如果没有执行工具调用，直接传递原始finish-step事件
+            controller.enqueue(chunk)
+
+            // 清理状态
+            textBuffer = ''
+            return
+          }
+
+          // 处理 finish 类型，使用累加后的 totalUsage
+          if (chunk.type === 'finish') {
+            controller.enqueue({
+              ...chunk,
+              totalUsage: context.accumulatedUsage
+            })
+            return
+          }
+
+          // 对于其他类型的事件，直接传递（不包括text-start，已在上面处理）
+          if ((chunk as any).type !== 'text-start') {
+            controller.enqueue(chunk)
+          }
+        },
+
+        flush() {
+          // 清理pending状态
+          pendingTextStart = null
+          hasStartedText = false
+        }
+      })
+    }
+  })
+}

packages/aiCore/src/core/plugins/built-in/toolUsePlugin/tagExtraction.ts

@@ -0,0 +1,196 @@
+// Copied from https://github.com/vercel/ai/blob/main/packages/ai/core/util/get-potential-start-index.ts
+
+/**
+ * Returns the index of the start of the searchedText in the text, or null if it
+ * is not found.
+ */
+export function getPotentialStartIndex(text: string, searchedText: string): number | null {
+  // Return null immediately if searchedText is empty.
+  if (searchedText.length === 0) {
+    return null
+  }
+
+  // Check if the searchedText exists as a direct substring of text.
+  const directIndex = text.indexOf(searchedText)
+  if (directIndex !== -1) {
+    return directIndex
+  }
+
+  // Otherwise, look for the largest suffix of "text" that matches
+  // a prefix of "searchedText". We go from the end of text inward.
+  for (let i = text.length - 1; i >= 0; i--) {
+    const suffix = text.substring(i)
+    if (searchedText.startsWith(suffix)) {
+      return i
+    }
+  }
+
+  return null
+}
+
+export interface TagConfig {
+  openingTag: string
+  closingTag: string
+  separator?: string
+}
+
+export interface TagExtractionState {
+  textBuffer: string
+  isInsideTag: boolean
+  isFirstTag: boolean
+  isFirstText: boolean
+  afterSwitch: boolean
+  accumulatedTagContent: string
+  hasTagContent: boolean
+}
+
+export interface TagExtractionResult {
+  content: string
+  isTagContent: boolean
+  complete: boolean
+  tagContentExtracted?: string
+}
+
+/**
+ * 通用标签提取处理器
+ * 可以处理各种形式的标签对，如 <think>...</think>, <tool_use>...</tool_use> 等
+ */
+export class TagExtractor {
+  private config: TagConfig
+  private state: TagExtractionState
+
+  constructor(config: TagConfig) {
+    this.config = config
+    this.state = {
+      textBuffer: '',
+      isInsideTag: false,
+      isFirstTag: true,
+      isFirstText: true,
+      afterSwitch: false,
+      accumulatedTagContent: '',
+      hasTagContent: false
+    }
+  }
+
+  /**
+   * 处理文本块，返回处理结果
+   */
+  processText(newText: string): TagExtractionResult[] {
+    this.state.textBuffer += newText
+    const results: TagExtractionResult[] = []
+
+    // 处理标签提取逻辑
+    while (true) {
+      const nextTag = this.state.isInsideTag ? this.config.closingTag : this.config.openingTag
+      const startIndex = getPotentialStartIndex(this.state.textBuffer, nextTag)
+
+      if (startIndex == null) {
+        const content = this.state.textBuffer
+        if (content.length > 0) {
+          results.push({
+            content: this.addPrefix(content),
+            isTagContent: this.state.isInsideTag,
+            complete: false
+          })
+
+          if (this.state.isInsideTag) {
+            this.state.accumulatedTagContent += this.addPrefix(content)
+            this.state.hasTagContent = true
+          }
+        }
+        this.state.textBuffer = ''
+        break
+      }
+
+      // 处理标签前的内容
+      const contentBeforeTag = this.state.textBuffer.slice(0, startIndex)
+      if (contentBeforeTag.length > 0) {
+        results.push({
+          content: this.addPrefix(contentBeforeTag),
+          isTagContent: this.state.isInsideTag,
+          complete: false
+        })
+
+        if (this.state.isInsideTag) {
+          this.state.accumulatedTagContent += this.addPrefix(contentBeforeTag)
+          this.state.hasTagContent = true
+        }
+      }
+
+      const foundFullMatch = startIndex + nextTag.length <= this.state.textBuffer.length
+
+      if (foundFullMatch) {
+        // 如果找到完整的标签
+        this.state.textBuffer = this.state.textBuffer.slice(startIndex + nextTag.length)
+
+        // 如果刚刚结束一个标签内容，生成完整的标签内容结果
+        if (this.state.isInsideTag && this.state.hasTagContent) {
+          results.push({
+            content: '',
+            isTagContent: false,
+            complete: true,
+            tagContentExtracted: this.state.accumulatedTagContent
+          })
+          this.state.accumulatedTagContent = ''
+          this.state.hasTagContent = false
+        }
+
+        this.state.isInsideTag = !this.state.isInsideTag
+        this.state.afterSwitch = true
+
+        if (this.state.isInsideTag) {
+          this.state.isFirstTag = false
+        } else {
+          this.state.isFirstText = false
+        }
+      } else {
+        this.state.textBuffer = this.state.textBuffer.slice(startIndex)
+        break
+      }
+    }
+
+    return results
+  }
+
+  /**
+   * 完成处理，返回任何剩余的标签内容
+   */
+  finalize(): TagExtractionResult | null {
+    if (this.state.hasTagContent && this.state.accumulatedTagContent) {
+      const result = {
+        content: '',
+        isTagContent: false,
+        complete: true,
+        tagContentExtracted: this.state.accumulatedTagContent
+      }
+      this.state.accumulatedTagContent = ''
+      this.state.hasTagContent = false
+      return result
+    }
+    return null
+  }
+
+  private addPrefix(text: string): string {
+    const needsPrefix =
+      this.state.afterSwitch && (this.state.isInsideTag ? !this.state.isFirstTag : !this.state.isFirstText)
+
+    const prefix = needsPrefix && this.config.separator ? this.config.separator : ''
+    this.state.afterSwitch = false
+    return prefix + text
+  }
+
+  /**
+   * 重置状态
+   */
+  reset(): void {
+    this.state = {
+      textBuffer: '',
+      isInsideTag: false,
+      isFirstTag: true,
+      isFirstText: true,
+      afterSwitch: false,
+      accumulatedTagContent: '',
+      hasTagContent: false
+    }
+  }
+}

packages/aiCore/src/core/plugins/built-in/toolUsePlugin/type.ts

@@ -0,0 +1,33 @@
+import { ToolSet } from 'ai'
+
+import { AiRequestContext } from '../..'
+
+/**
+ * 解析结果类型
+ * 表示从AI响应中解析出的工具使用意图
+ */
+export interface ToolUseResult {
+  id: string
+  toolName: string
+  arguments: any
+  status: 'pending' | 'invoking' | 'done' | 'error'
+}
+
+export interface BaseToolUsePluginConfig {
+  enabled?: boolean
+}
+
+export interface PromptToolUseConfig extends BaseToolUsePluginConfig {
+  // 自定义系统提示符构建函数（可选，有默认实现）
+  buildSystemPrompt?: (userSystemPrompt: string, tools: ToolSet) => string
+  // 自定义工具解析函数（可选，有默认实现）
+  parseToolUse?: (content: string, tools: ToolSet) => { results: ToolUseResult[]; content: string }
+  createSystemMessage?: (systemPrompt: string, originalParams: any, context: AiRequestContext) => string | null
+}
+
+/**
+ * 扩展的 AI 请求上下文，支持 MCP 工具存储
+ */
+export interface ToolUseRequestContext extends AiRequestContext {
+  mcpTools: ToolSet
+}

packages/aiCore/src/core/plugins/built-in/webSearchPlugin/helper.ts

@@ -0,0 +1,81 @@
+import { anthropic } from '@ai-sdk/anthropic'
+import { google } from '@ai-sdk/google'
+import { openai } from '@ai-sdk/openai'
+
+import { ProviderOptionsMap } from '../../../options/types'
+import { OpenRouterSearchConfig } from './openrouter'
+
+/**
+ * 从 AI SDK 的工具函数中提取参数类型，以确保类型安全。
+ */
+export type OpenAISearchConfig = NonNullable<Parameters<typeof openai.tools.webSearch>[0]>
+export type OpenAISearchPreviewConfig = NonNullable<Parameters<typeof openai.tools.webSearchPreview>[0]>
+export type AnthropicSearchConfig = NonNullable<Parameters<typeof anthropic.tools.webSearch_20250305>[0]>
+export type GoogleSearchConfig = NonNullable<Parameters<typeof google.tools.googleSearch>[0]>
+export type XAISearchConfig = NonNullable<ProviderOptionsMap['xai']['searchParameters']>
+
+/**
+ * 插件初始化时接收的完整配置对象
+ *
+ * 其结构与 ProviderOptions 保持一致，方便上游统一管理配置
+ */
+export interface WebSearchPluginConfig {
+  openai?: OpenAISearchConfig
+  'openai-chat'?: OpenAISearchPreviewConfig
+  anthropic?: AnthropicSearchConfig
+  xai?: ProviderOptionsMap['xai']['searchParameters']
+  google?: GoogleSearchConfig
+  'google-vertex'?: GoogleSearchConfig
+  openrouter?: OpenRouterSearchConfig
+}
+
+/**
+ * 插件的默认配置
+ */
+export const DEFAULT_WEB_SEARCH_CONFIG: WebSearchPluginConfig = {
+  google: {},
+  'google-vertex': {},
+  openai: {},
+  'openai-chat': {},
+  xai: {
+    mode: 'on',
+    returnCitations: true,
+    maxSearchResults: 5,
+    sources: [{ type: 'web' }, { type: 'x' }, { type: 'news' }]
+  },
+  anthropic: {
+    maxUses: 5
+  },
+  openrouter: {
+    plugins: [
+      {
+        id: 'web',
+        max_results: 5
+      }
+    ]
+  }
+}
+
+export type WebSearchToolOutputSchema = {
+  // Anthropic 工具 - 手动定义
+  anthropicWebSearch: Array<{
+    url: string
+    title: string
+    pageAge: string | null
+    encryptedContent: string
+    type: string
+  }>
+
+  // OpenAI 工具 - 基于实际输出
+  openaiWebSearch: {
+    status: 'completed' | 'failed'
+  }
+
+  // Google 工具
+  googleSearch: {
+    webSearchQueries?: string[]
+    groundingChunks?: Array<{
+      web?: { uri: string; title: string }
+    }>
+  }
+}

packages/aiCore/src/core/plugins/built-in/webSearchPlugin/index.ts

@@ -0,0 +1,84 @@
+/**
+ * Web Search Plugin
+ * 提供统一的网络搜索能力，支持多个 AI Provider
+ */
+import { anthropic } from '@ai-sdk/anthropic'
+import { google } from '@ai-sdk/google'
+import { openai } from '@ai-sdk/openai'
+
+import { createOpenRouterOptions, createXaiOptions, mergeProviderOptions } from '../../../options'
+import { definePlugin } from '../../'
+import type { AiRequestContext } from '../../types'
+import { DEFAULT_WEB_SEARCH_CONFIG, WebSearchPluginConfig } from './helper'
+
+/**
+ * 网络搜索插件
+ *
+ * @param config - 在插件初始化时传入的静态配置
+ */
+export const webSearchPlugin = (config: WebSearchPluginConfig = DEFAULT_WEB_SEARCH_CONFIG) =>
+  definePlugin({
+    name: 'webSearch',
+    enforce: 'pre',
+
+    transformParams: async (params: any, context: AiRequestContext) => {
+      const { providerId } = context
+      switch (providerId) {
+        case 'openai': {
+          if (config.openai) {
+            if (!params.tools) params.tools = {}
+            params.tools.web_search = openai.tools.webSearch(config.openai)
+          }
+          break
+        }
+        case 'openai-chat': {
+          if (config['openai-chat']) {
+            if (!params.tools) params.tools = {}
+            params.tools.web_search_preview = openai.tools.webSearchPreview(config['openai-chat'])
+          }
+          break
+        }
+
+        case 'anthropic': {
+          if (config.anthropic) {
+            if (!params.tools) params.tools = {}
+            params.tools.web_search = anthropic.tools.webSearch_20250305(config.anthropic)
+          }
+          break
+        }
+
+        case 'google': {
+          // case 'google-vertex':
+          if (!params.tools) params.tools = {}
+          params.tools.web_search = google.tools.googleSearch(config.google || {})
+          break
+        }
+
+        case 'xai': {
+          if (config.xai) {
+            const searchOptions = createXaiOptions({
+              searchParameters: { ...config.xai, mode: 'on' }
+            })
+            params.providerOptions = mergeProviderOptions(params.providerOptions, searchOptions)
+          }
+          break
+        }
+
+        case 'openrouter': {
+          if (config.openrouter) {
+            const searchOptions = createOpenRouterOptions(config.openrouter)
+            params.providerOptions = mergeProviderOptions(params.providerOptions, searchOptions)
+          }
+          break
+        }
+      }
+
+      return params
+    }
+  })
+
+// 导出类型定义供开发者使用
+export type { WebSearchPluginConfig, WebSearchToolOutputSchema } from './helper'
+
+// 默认导出
+export default webSearchPlugin

packages/aiCore/src/core/plugins/built-in/webSearchPlugin/openrouter.ts

@@ -0,0 +1,26 @@
+export type OpenRouterSearchConfig = {
+  plugins?: Array<{
+    id: 'web'
+    /**
+     * Maximum number of search results to include (default: 5)
+     */
+    max_results?: number
+    /**
+     * Custom search prompt to guide the search query
+     */
+    search_prompt?: string
+  }>
+  /**
+   * Built-in web search options for models that support native web search
+   */
+  web_search_options?: {
+    /**
+     * Maximum number of search results to include
+     */
+    max_results?: number
+    /**
+     * Custom search prompt to guide the search query
+     */
+    search_prompt?: string
+  }
+}

packages/aiCore/src/core/plugins/index.ts

@@ -0,0 +1,35 @@
+// 核心类型和接口
+export type { AiPlugin, AiRequestContext, HookResult, PluginManagerConfig } from './types'
+import type { ImageModelV2 } from '@ai-sdk/provider'
+import type { LanguageModel } from 'ai'
+
+import type { ProviderId } from '../providers'
+import type { AiPlugin, AiRequestContext } from './types'
+
+// 插件管理器
+export { PluginManager } from './manager'
+
+// 工具函数
+export function createContext<T extends ProviderId>(
+  providerId: T,
+  model: LanguageModel | ImageModelV2,
+  originalParams: any
+): AiRequestContext {
+  return {
+    providerId,
+    model,
+    originalParams,
+    metadata: {},
+    startTime: Date.now(),
+    requestId: `${providerId}-${typeof model === 'string' ? model : model?.modelId}-${Date.now()}-${Math.random().toString(36).slice(2)}`,
+    // 占位
+    recursiveCall: () => Promise.resolve(null)
+  }
+}
+
+// 插件构建器 - 便于创建插件
+export function definePlugin(plugin: AiPlugin): AiPlugin
+export function definePlugin<T extends (...args: any[]) => AiPlugin>(pluginFactory: T): T
+export function definePlugin(plugin: AiPlugin | ((...args: any[]) => AiPlugin)) {
+  return plugin
+}

packages/aiCore/src/core/plugins/manager.ts

@@ -0,0 +1,184 @@
+import { AiPlugin, AiRequestContext } from './types'
+
+/**
+ * 插件管理器
+ */
+export class PluginManager {
+  private plugins: AiPlugin[] = []
+
+  constructor(plugins: AiPlugin[] = []) {
+    this.plugins = this.sortPlugins(plugins)
+  }
+
+  /**
+   * 添加插件
+   */
+  use(plugin: AiPlugin): this {
+    this.plugins = this.sortPlugins([...this.plugins, plugin])
+    return this
+  }
+
+  /**
+   * 移除插件
+   */
+  remove(pluginName: string): this {
+    this.plugins = this.plugins.filter((p) => p.name !== pluginName)
+    return this
+  }
+
+  /**
+   * 插件排序：pre -> normal -> post
+   */
+  private sortPlugins(plugins: AiPlugin[]): AiPlugin[] {
+    const pre: AiPlugin[] = []
+    const normal: AiPlugin[] = []
+    const post: AiPlugin[] = []
+
+    plugins.forEach((plugin) => {
+      if (plugin.enforce === 'pre') {
+        pre.push(plugin)
+      } else if (plugin.enforce === 'post') {
+        post.push(plugin)
+      } else {
+        normal.push(plugin)
+      }
+    })
+
+    return [...pre, ...normal, ...post]
+  }
+
+  /**
+   * 执行 First 钩子 - 返回第一个有效结果
+   */
+  async executeFirst<T>(
+    hookName: 'resolveModel' | 'loadTemplate',
+    arg: any,
+    context: AiRequestContext
+  ): Promise<T | null> {
+    for (const plugin of this.plugins) {
+      const hook = plugin[hookName]
+      if (hook) {
+        const result = await hook(arg, context)
+        if (result !== null && result !== undefined) {
+          return result as T
+        }
+      }
+    }
+    return null
+  }
+
+  /**
+   * 执行 Sequential 钩子 - 链式数据转换
+   */
+  async executeSequential<T>(
+    hookName: 'transformParams' | 'transformResult',
+    initialValue: T,
+    context: AiRequestContext
+  ): Promise<T> {
+    let result = initialValue
+
+    for (const plugin of this.plugins) {
+      const hook = plugin[hookName]
+      if (hook) {
+        result = await hook<T>(result, context)
+      }
+    }
+
+    return result
+  }
+
+  /**
+   * 执行 ConfigureContext 钩子 - 串行配置上下文
+   */
+  async executeConfigureContext(context: AiRequestContext): Promise<void> {
+    for (const plugin of this.plugins) {
+      const hook = plugin.configureContext
+      if (hook) {
+        await hook(context)
+      }
+    }
+  }
+
+  /**
+   * 执行 Parallel 钩子 - 并行副作用
+   */
+  async executeParallel(
+    hookName: 'onRequestStart' | 'onRequestEnd' | 'onError',
+    context: AiRequestContext,
+    result?: any,
+    error?: Error
+  ): Promise<void> {
+    const promises = this.plugins
+      .map((plugin) => {
+        const hook = plugin[hookName]
+        if (!hook) return null
+
+        if (hookName === 'onError' && error) {
+          return (hook as any)(error, context)
+        } else if (hookName === 'onRequestEnd' && result !== undefined) {
+          return (hook as any)(context, result)
+        } else if (hookName === 'onRequestStart') {
+          return (hook as any)(context)
+        }
+        return null
+      })
+      .filter(Boolean)
+
+    // 使用 Promise.all 而不是 allSettled，让插件错误能够抛出
+    await Promise.all(promises)
+  }
+
+  /**
+   * 收集所有流转换器（返回数组，AI SDK 原生支持）
+   */
+  collectStreamTransforms(params: any, context: AiRequestContext) {
+    return this.plugins
+      .filter((plugin) => plugin.transformStream)
+      .map((plugin) => plugin.transformStream?.(params, context))
+  }
+
+  /**
+   * 获取所有插件信息
+   */
+  getPlugins(): AiPlugin[] {
+    return [...this.plugins]
+  }
+
+  /**
+   * 获取插件统计信息
+   */
+  getStats() {
+    const stats = {
+      total: this.plugins.length,
+      pre: 0,
+      normal: 0,
+      post: 0,
+      hooks: {
+        resolveModel: 0,
+        loadTemplate: 0,
+        transformParams: 0,
+        transformResult: 0,
+        onRequestStart: 0,
+        onRequestEnd: 0,
+        onError: 0,
+        transformStream: 0
+      }
+    }
+
+    this.plugins.forEach((plugin) => {
+      // 统计 enforce 类型
+      if (plugin.enforce === 'pre') stats.pre++
+      else if (plugin.enforce === 'post') stats.post++
+      else stats.normal++
+
+      // 统计钩子数量
+      Object.keys(stats.hooks).forEach((hookName) => {
+        if (plugin[hookName as keyof AiPlugin]) {
+          stats.hooks[hookName as keyof typeof stats.hooks]++
+        }
+      })
+    })
+
+    return stats
+  }
+}

packages/aiCore/src/core/plugins/types.ts

@@ -0,0 +1,79 @@
+import type { ImageModelV2 } from '@ai-sdk/provider'
+import type { LanguageModel, TextStreamPart, ToolSet } from 'ai'
+
+import { type ProviderId } from '../providers/types'
+
+/**
+ * 递归调用函数类型
+ * 使用 any 是因为递归调用时参数和返回类型可能完全不同
+ */
+export type RecursiveCallFn = (newParams: any) => Promise<any>
+
+/**
+ * AI 请求上下文
+ */
+export interface AiRequestContext {
+  providerId: ProviderId
+  model: LanguageModel | ImageModelV2
+  originalParams: any
+  metadata: Record<string, any>
+  startTime: number
+  requestId: string
+  recursiveCall: RecursiveCallFn
+  isRecursiveCall?: boolean
+  mcpTools?: ToolSet
+  [key: string]: any
+}
+
+/**
+ * 钩子分类
+ */
+export interface AiPlugin {
+  name: string
+  enforce?: 'pre' | 'post'
+
+  // 【First】首个钩子 - 只执行第一个返回值的插件
+  resolveModel?: (
+    modelId: string,
+    context: AiRequestContext
+  ) => Promise<LanguageModel | ImageModelV2 | null> | LanguageModel | ImageModelV2 | null
+  loadTemplate?: (templateName: string, context: AiRequestContext) => any | null | Promise<any | null>
+
+  // 【Sequential】串行钩子 - 链式执行，支持数据转换
+  configureContext?: (context: AiRequestContext) => void | Promise<void>
+  transformParams?: <T>(params: T, context: AiRequestContext) => T | Promise<T>
+  transformResult?: <T>(result: T, context: AiRequestContext) => T | Promise<T>
+
+  // 【Parallel】并行钩子 - 不依赖顺序，用于副作用
+  onRequestStart?: (context: AiRequestContext) => void | Promise<void>
+  onRequestEnd?: (context: AiRequestContext, result: any) => void | Promise<void>
+  onError?: (error: Error, context: AiRequestContext) => void | Promise<void>
+
+  // 【Stream】流处理 - 直接使用 AI SDK
+  transformStream?: (
+    params: any,
+    context: AiRequestContext
+  ) => <TOOLS extends ToolSet>(options?: {
+    tools: TOOLS
+    stopStream: () => void
+  }) => TransformStream<TextStreamPart<TOOLS>, TextStreamPart<TOOLS>>
+
+  // AI SDK 原生中间件
+  // aiSdkMiddlewares?: LanguageModelV1Middleware[]
+}
+
+/**
+ * 插件管理器配置
+ */
+export interface PluginManagerConfig {
+  plugins: AiPlugin[]
+  context: Partial<AiRequestContext>
+}
+
+/**
+ * 钩子执行结果
+ */
+export interface HookResult<T = any> {
+  value: T
+  stop?: boolean
+}

packages/aiCore/src/core/providers/HubProvider.ts

@@ -0,0 +1,101 @@
+/**
+ * Hub Provider - 支持路由到多个底层provider
+ *
+ * 支持格式: hubId:providerId:modelId
+ * 例如: aihubmix:anthropic:claude-3.5-sonnet
+ */
+
+import { ProviderV2 } from '@ai-sdk/provider'
+import { customProvider } from 'ai'
+
+import { globalRegistryManagement } from './RegistryManagement'
+import type { AiSdkMethodName, AiSdkModelReturn, AiSdkModelType } from './types'
+
+export interface HubProviderConfig {
+  /** Hub的唯一标识符 */
+  hubId: string
+  /** 是否启用调试日志 */
+  debug?: boolean
+}
+
+export class HubProviderError extends Error {
+  constructor(
+    message: string,
+    public readonly hubId: string,
+    public readonly providerId?: string,
+    public readonly originalError?: Error
+  ) {
+    super(message)
+    this.name = 'HubProviderError'
+  }
+}
+
+/**
+ * 解析Hub模型ID
+ */
+function parseHubModelId(modelId: string): { provider: string; actualModelId: string } {
+  const parts = modelId.split(':')
+  if (parts.length !== 2) {
+    throw new HubProviderError(`Invalid hub model ID format. Expected "provider:modelId", got: ${modelId}`, 'unknown')
+  }
+  return {
+    provider: parts[0],
+    actualModelId: parts[1]
+  }
+}
+
+/**
+ * 创建Hub Provider
+ */
+export function createHubProvider(config: HubProviderConfig): ProviderV2 {
+  const { hubId } = config
+
+  function getTargetProvider(providerId: string): ProviderV2 {
+    // 从全局注册表获取provider实例
+    try {
+      const provider = globalRegistryManagement.getProvider(providerId)
+      if (!provider) {
+        throw new HubProviderError(
+          `Provider "${providerId}" is not initialized. Please call initializeProvider("${providerId}", options) first.`,
+          hubId,
+          providerId
+        )
+      }
+      return provider
+    } catch (error) {
+      throw new HubProviderError(
+        `Failed to get provider "${providerId}": ${error instanceof Error ? error.message : 'Unknown error'}`,
+        hubId,
+        providerId,
+        error instanceof Error ? error : undefined
+      )
+    }
+  }
+
+  function resolveModel<T extends AiSdkModelType>(
+    modelId: string,
+    modelType: T,
+    methodName: AiSdkMethodName<T>
+  ): AiSdkModelReturn<T> {
+    const { provider, actualModelId } = parseHubModelId(modelId)
+    const targetProvider = getTargetProvider(provider)
+
+    const fn = targetProvider[methodName] as (id: string) => AiSdkModelReturn<T>
+
+    if (!fn) {
+      throw new HubProviderError(`Provider "${provider}" does not support ${modelType}`, hubId, provider)
+    }
+
+    return fn(actualModelId)
+  }
+
+  return customProvider({
+    fallbackProvider: {
+      languageModel: (modelId: string) => resolveModel(modelId, 'text', 'languageModel'),
+      textEmbeddingModel: (modelId: string) => resolveModel(modelId, 'embedding', 'textEmbeddingModel'),
+      imageModel: (modelId: string) => resolveModel(modelId, 'image', 'imageModel'),
+      transcriptionModel: (modelId: string) => resolveModel(modelId, 'transcription', 'transcriptionModel'),
+      speechModel: (modelId: string) => resolveModel(modelId, 'speech', 'speechModel')
+    }
+  })
+}

packages/aiCore/src/core/providers/RegistryManagement.ts

@@ -0,0 +1,221 @@
+/**
+ * Provider 注册表管理器
+ * 纯粹的管理功能：存储、检索已配置好的 provider 实例
+ * 基于 AI SDK 原生的 createProviderRegistry
+ */
+
+import { EmbeddingModelV2, ImageModelV2, LanguageModelV2, ProviderV2 } from '@ai-sdk/provider'
+import { createProviderRegistry, type ProviderRegistryProvider } from 'ai'
+
+type PROVIDERS = Record<string, ProviderV2>
+
+export const DEFAULT_SEPARATOR = '|'
+
+// export type MODEL_ID = `${string}${typeof DEFAULT_SEPARATOR}${string}`
+
+export class RegistryManagement<SEPARATOR extends string = typeof DEFAULT_SEPARATOR> {
+  private providers: PROVIDERS = {}
+  private aliases: Set<string> = new Set() // 记录哪些key是别名
+  private separator: SEPARATOR
+  private registry: ProviderRegistryProvider<PROVIDERS, SEPARATOR> | null = null
+
+  constructor(options: { separator: SEPARATOR } = { separator: DEFAULT_SEPARATOR as SEPARATOR }) {
+    this.separator = options.separator
+  }
+
+  /**
+   * 注册已配置好的 provider 实例
+   */
+  registerProvider(id: string, provider: ProviderV2, aliases?: string[]): this {
+    // 注册主provider
+    this.providers[id] = provider
+
+    // 注册别名（都指向同一个provider实例）
+    if (aliases) {
+      aliases.forEach((alias) => {
+        this.providers[alias] = provider // 直接存储引用
+        this.aliases.add(alias) // 标记为别名
+      })
+    }
+
+    this.rebuildRegistry()
+    return this
+  }
+
+  /**
+   * 获取已注册的provider实例
+   */
+  getProvider(id: string): ProviderV2 | undefined {
+    return this.providers[id]
+  }
+
+  /**
+   * 批量注册 providers
+   */
+  registerProviders(providers: Record<string, ProviderV2>): this {
+    Object.assign(this.providers, providers)
+    this.rebuildRegistry()
+    return this
+  }
+
+  /**
+   * 移除 provider（同时清理相关别名）
+   */
+  unregisterProvider(id: string): this {
+    const provider = this.providers[id]
+    if (!provider) return this
+
+    // 如果移除的是真实ID，需要清理所有指向它的别名
+    if (!this.aliases.has(id)) {
+      // 找到所有指向此provider的别名并删除
+      const aliasesToRemove: string[] = []
+      this.aliases.forEach((alias) => {
+        if (this.providers[alias] === provider) {
+          aliasesToRemove.push(alias)
+        }
+      })
+
+      aliasesToRemove.forEach((alias) => {
+        delete this.providers[alias]
+        this.aliases.delete(alias)
+      })
+    } else {
+      // 如果移除的是别名，只删除别名记录
+      this.aliases.delete(id)
+    }
+
+    delete this.providers[id]
+    this.rebuildRegistry()
+    return this
+  }
+
+  /**
+   * 立即重建 registry - 每次变更都重建
+   */
+  private rebuildRegistry(): void {
+    if (Object.keys(this.providers).length === 0) {
+      this.registry = null
+      return
+    }
+
+    this.registry = createProviderRegistry<PROVIDERS, SEPARATOR>(this.providers, {
+      separator: this.separator
+    })
+  }
+
+  /**
+   * 获取语言模型 - AI SDK 原生方法
+   */
+  languageModel(id: `${string}${SEPARATOR}${string}`): LanguageModelV2 {
+    if (!this.registry) {
+      throw new Error('No providers registered')
+    }
+    return this.registry.languageModel(id)
+  }
+
+  /**
+   * 获取文本嵌入模型 - AI SDK 原生方法
+   */
+  textEmbeddingModel(id: `${string}${SEPARATOR}${string}`): EmbeddingModelV2<string> {
+    if (!this.registry) {
+      throw new Error('No providers registered')
+    }
+    return this.registry.textEmbeddingModel(id)
+  }
+
+  /**
+   * 获取图像模型 - AI SDK 原生方法
+   */
+  imageModel(id: `${string}${SEPARATOR}${string}`): ImageModelV2 {
+    if (!this.registry) {
+      throw new Error('No providers registered')
+    }
+    return this.registry.imageModel(id)
+  }
+
+  /**
+   * 获取转录模型 - AI SDK 原生方法
+   */
+  transcriptionModel(id: `${string}${SEPARATOR}${string}`): any {
+    if (!this.registry) {
+      throw new Error('No providers registered')
+    }
+    return this.registry.transcriptionModel(id)
+  }
+
+  /**
+   * 获取语音模型 - AI SDK 原生方法
+   */
+  speechModel(id: `${string}${SEPARATOR}${string}`): any {
+    if (!this.registry) {
+      throw new Error('No providers registered')
+    }
+    return this.registry.speechModel(id)
+  }
+
+  /**
+   * 获取已注册的 provider 列表
+   */
+  getRegisteredProviders(): string[] {
+    return Object.keys(this.providers)
+  }
+
+  /**
+   * 检查是否有已注册的 providers
+   */
+  hasProviders(): boolean {
+    return Object.keys(this.providers).length > 0
+  }
+
+  /**
+   * 清除所有 providers
+   */
+  clear(): this {
+    this.providers = {}
+    this.aliases.clear()
+    this.registry = null
+    return this
+  }
+
+  /**
+   * 解析真实的Provider ID（供getAiSdkProviderId使用）
+   * 如果传入的是别名，返回真实的Provider ID
+   * 如果传入的是真实ID，直接返回
+   */
+  resolveProviderId(id: string): string {
+    if (!this.aliases.has(id)) return id // 不是别名，直接返回
+
+    // 是别名，找到真实ID
+    const targetProvider = this.providers[id]
+    for (const [realId, provider] of Object.entries(this.providers)) {
+      if (provider === targetProvider && !this.aliases.has(realId)) {
+        return realId
+      }
+    }
+    return id
+  }
+
+  /**
+   * 检查是否为别名
+   */
+  isAlias(id: string): boolean {
+    return this.aliases.has(id)
+  }
+
+  /**
+   * 获取所有别名映射关系
+   */
+  getAllAliases(): Record<string, string> {
+    const result: Record<string, string> = {}
+    this.aliases.forEach((alias) => {
+      result[alias] = this.resolveProviderId(alias)
+    })
+    return result
+  }
+}
+
+/**
+ * 全局注册表管理器实例
+ * 使用 | 作为分隔符，因为 : 会和 :free 等suffix冲突
+ */
+export const globalRegistryManagement = new RegistryManagement()

packages/aiCore/src/core/providers/__tests__/registry-functionality.test.ts

@@ -0,0 +1,632 @@
+/**
+ * 测试真正的 AiProviderRegistry 功能
+ */
+
+import { beforeEach, describe, expect, it, vi } from 'vitest'
+
+// 模拟 AI SDK
+vi.mock('@ai-sdk/openai', () => ({
+  createOpenAI: vi.fn(() => ({ name: 'openai-mock' }))
+}))
+
+vi.mock('@ai-sdk/anthropic', () => ({
+  createAnthropic: vi.fn(() => ({ name: 'anthropic-mock' }))
+}))
+
+vi.mock('@ai-sdk/azure', () => ({
+  createAzure: vi.fn(() => ({ name: 'azure-mock' }))
+}))
+
+vi.mock('@ai-sdk/deepseek', () => ({
+  createDeepSeek: vi.fn(() => ({ name: 'deepseek-mock' }))
+}))
+
+vi.mock('@ai-sdk/google', () => ({
+  createGoogleGenerativeAI: vi.fn(() => ({ name: 'google-mock' }))
+}))
+
+vi.mock('@ai-sdk/openai-compatible', () => ({
+  createOpenAICompatible: vi.fn(() => ({ name: 'openai-compatible-mock' }))
+}))
+
+vi.mock('@ai-sdk/xai', () => ({
+  createXai: vi.fn(() => ({ name: 'xai-mock' }))
+}))
+
+import {
+  cleanup,
+  clearAllProviders,
+  createAndRegisterProvider,
+  createProvider,
+  getAllProviderConfigAliases,
+  getAllProviderConfigs,
+  getInitializedProviders,
+  getLanguageModel,
+  getProviderConfig,
+  getProviderConfigByAlias,
+  getSupportedProviders,
+  hasInitializedProviders,
+  hasProviderConfig,
+  hasProviderConfigByAlias,
+  isProviderConfigAlias,
+  ProviderInitializationError,
+  providerRegistry,
+  registerMultipleProviderConfigs,
+  registerProvider,
+  registerProviderConfig,
+  resolveProviderConfigId
+} from '../registry'
+import type { ProviderConfig } from '../schemas'
+
+describe('Provider Registry 功能测试', () => {
+  beforeEach(() => {
+    // 清理状态
+    cleanup()
+  })
+
+  describe('基础功能', () => {
+    it('能够获取支持的 providers 列表', () => {
+      const providers = getSupportedProviders()
+      expect(Array.isArray(providers)).toBe(true)
+      expect(providers.length).toBeGreaterThan(0)
+
+      // 检查返回的数据结构
+      providers.forEach((provider) => {
+        expect(provider).toHaveProperty('id')
+        expect(provider).toHaveProperty('name')
+        expect(typeof provider.id).toBe('string')
+        expect(typeof provider.name).toBe('string')
+      })
+
+      // 包含基础 providers
+      const providerIds = providers.map((p) => p.id)
+      expect(providerIds).toContain('openai')
+      expect(providerIds).toContain('anthropic')
+      expect(providerIds).toContain('google')
+    })
+
+    it('能够获取已初始化的 providers', () => {
+      // 初始状态下没有已初始化的 providers
+      expect(getInitializedProviders()).toEqual([])
+      expect(hasInitializedProviders()).toBe(false)
+    })
+
+    it('能够访问全局注册管理器', () => {
+      expect(providerRegistry).toBeDefined()
+      expect(typeof providerRegistry.clear).toBe('function')
+      expect(typeof providerRegistry.getRegisteredProviders).toBe('function')
+      expect(typeof providerRegistry.hasProviders).toBe('function')
+    })
+
+    it('能够获取语言模型', () => {
+      // 在没有注册 provider 的情况下，这个函数应该会抛出错误
+      expect(() => getLanguageModel('non-existent')).toThrow('No providers registered')
+    })
+  })
+
+  describe('Provider 配置注册', () => {
+    it('能够注册自定义 provider 配置', () => {
+      const config: ProviderConfig = {
+        id: 'custom-provider',
+        name: 'Custom Provider',
+        creator: vi.fn(() => ({ name: 'custom' })),
+        supportsImageGeneration: false
+      }
+
+      const success = registerProviderConfig(config)
+      expect(success).toBe(true)
+
+      expect(hasProviderConfig('custom-provider')).toBe(true)
+      expect(getProviderConfig('custom-provider')).toEqual(config)
+    })
+
+    it('能够注册带别名的 provider 配置', () => {
+      const config: ProviderConfig = {
+        id: 'custom-provider-with-aliases',
+        name: 'Custom Provider with Aliases',
+        creator: vi.fn(() => ({ name: 'custom-aliased' })),
+        supportsImageGeneration: false,
+        aliases: ['alias-1', 'alias-2']
+      }
+
+      const success = registerProviderConfig(config)
+      expect(success).toBe(true)
+
+      expect(hasProviderConfigByAlias('alias-1')).toBe(true)
+      expect(hasProviderConfigByAlias('alias-2')).toBe(true)
+      expect(getProviderConfigByAlias('alias-1')).toEqual(config)
+      expect(resolveProviderConfigId('alias-1')).toBe('custom-provider-with-aliases')
+    })
+
+    it('拒绝无效的配置', () => {
+      // 缺少必要字段
+      const invalidConfig = {
+        id: 'invalid-provider'
+        // 缺少 name, creator 等
+      }
+
+      const success = registerProviderConfig(invalidConfig as any)
+      expect(success).toBe(false)
+    })
+
+    it('能够批量注册 provider 配置', () => {
+      const configs: ProviderConfig[] = [
+        {
+          id: 'provider-1',
+          name: 'Provider 1',
+          creator: vi.fn(() => ({ name: 'provider-1' })),
+          supportsImageGeneration: false
+        },
+        {
+          id: 'provider-2',
+          name: 'Provider 2',
+          creator: vi.fn(() => ({ name: 'provider-2' })),
+          supportsImageGeneration: true
+        },
+        {
+          id: '', // 无效配置
+          name: 'Invalid Provider',
+          creator: vi.fn(() => ({ name: 'invalid' })),
+          supportsImageGeneration: false
+        } as any
+      ]
+
+      const successCount = registerMultipleProviderConfigs(configs)
+      expect(successCount).toBe(2) // 只有前两个成功
+
+      expect(hasProviderConfig('provider-1')).toBe(true)
+      expect(hasProviderConfig('provider-2')).toBe(true)
+      expect(hasProviderConfig('')).toBe(false)
+    })
+
+    it('能够获取所有配置和别名信息', () => {
+      // 注册一些配置
+      registerProviderConfig({
+        id: 'test-provider',
+        name: 'Test Provider',
+        creator: vi.fn(),
+        supportsImageGeneration: false,
+        aliases: ['test-alias']
+      })
+
+      const allConfigs = getAllProviderConfigs()
+      expect(Array.isArray(allConfigs)).toBe(true)
+      expect(allConfigs.some((config) => config.id === 'test-provider')).toBe(true)
+
+      const aliases = getAllProviderConfigAliases()
+      expect(aliases['test-alias']).toBe('test-provider')
+      expect(isProviderConfigAlias('test-alias')).toBe(true)
+    })
+  })
+
+  describe('Provider 创建和注册', () => {
+    it('能够创建 provider 实例', async () => {
+      const config: ProviderConfig = {
+        id: 'test-create-provider',
+        name: 'Test Create Provider',
+        creator: vi.fn(() => ({ name: 'test-created' })),
+        supportsImageGeneration: false
+      }
+
+      // 先注册配置
+      registerProviderConfig(config)
+
+      // 创建 provider 实例
+      const provider = await createProvider('test-create-provider', { apiKey: 'test' })
+      expect(provider).toBeDefined()
+      expect(config.creator).toHaveBeenCalledWith({ apiKey: 'test' })
+    })
+
+    it('能够注册 provider 到全局管理器', () => {
+      const mockProvider = { name: 'mock-provider' }
+      const config: ProviderConfig = {
+        id: 'test-register-provider',
+        name: 'Test Register Provider',
+        creator: vi.fn(() => mockProvider),
+        supportsImageGeneration: false
+      }
+
+      // 先注册配置
+      registerProviderConfig(config)
+
+      // 注册 provider 到全局管理器
+      const success = registerProvider('test-register-provider', mockProvider)
+      expect(success).toBe(true)
+
+      // 验证注册成功
+      const registeredProviders = getInitializedProviders()
+      expect(registeredProviders).toContain('test-register-provider')
+      expect(hasInitializedProviders()).toBe(true)
+    })
+
+    it('能够一步完成创建和注册', async () => {
+      const config: ProviderConfig = {
+        id: 'test-create-and-register',
+        name: 'Test Create and Register',
+        creator: vi.fn(() => ({ name: 'test-both' })),
+        supportsImageGeneration: false
+      }
+
+      // 先注册配置
+      registerProviderConfig(config)
+
+      // 一步完成创建和注册
+      const success = await createAndRegisterProvider('test-create-and-register', { apiKey: 'test' })
+      expect(success).toBe(true)
+
+      // 验证注册成功
+      const registeredProviders = getInitializedProviders()
+      expect(registeredProviders).toContain('test-create-and-register')
+    })
+  })
+
+  describe('Registry 管理', () => {
+    it('能够清理所有配置和注册的 providers', () => {
+      // 注册一些配置
+      registerProviderConfig({
+        id: 'temp-provider',
+        name: 'Temp Provider',
+        creator: vi.fn(() => ({ name: 'temp' })),
+        supportsImageGeneration: false
+      })
+
+      expect(hasProviderConfig('temp-provider')).toBe(true)
+
+      // 清理
+      cleanup()
+
+      expect(hasProviderConfig('temp-provider')).toBe(false)
+      // 但基础配置应该重新加载
+      expect(hasProviderConfig('openai')).toBe(true) // 基础 providers 会重新初始化
+    })
+
+    it('能够单独清理已注册的 providers', () => {
+      // 清理所有 providers
+      clearAllProviders()
+
+      expect(getInitializedProviders()).toEqual([])
+      expect(hasInitializedProviders()).toBe(false)
+    })
+
+    it('ProviderInitializationError 错误类工作正常', () => {
+      const error = new ProviderInitializationError('Test error', 'test-provider')
+      expect(error.message).toBe('Test error')
+      expect(error.providerId).toBe('test-provider')
+      expect(error.name).toBe('ProviderInitializationError')
+    })
+  })
+
+  describe('错误处理', () => {
+    it('优雅处理空配置', () => {
+      const success = registerProviderConfig(null as any)
+      expect(success).toBe(false)
+    })
+
+    it('优雅处理未定义配置', () => {
+      const success = registerProviderConfig(undefined as any)
+      expect(success).toBe(false)
+    })
+
+    it('处理空字符串 ID', () => {
+      const config = {
+        id: '',
+        name: 'Empty ID Provider',
+        creator: vi.fn(() => ({ name: 'empty' })),
+        supportsImageGeneration: false
+      }
+
+      const success = registerProviderConfig(config)
+      expect(success).toBe(false)
+    })
+
+    it('处理创建不存在配置的 provider', async () => {
+      await expect(createProvider('non-existent-provider', {})).rejects.toThrow(
+        'ProviderConfig not found for id: non-existent-provider'
+      )
+    })
+
+    it('处理注册不存在配置的 provider', () => {
+      const mockProvider = { name: 'mock' }
+      const success = registerProvider('non-existent-provider', mockProvider)
+      expect(success).toBe(false)
+    })
+
+    it('处理获取不存在配置的情况', () => {
+      expect(getProviderConfig('non-existent')).toBeUndefined()
+      expect(getProviderConfigByAlias('non-existent-alias')).toBeUndefined()
+      expect(hasProviderConfig('non-existent')).toBe(false)
+      expect(hasProviderConfigByAlias('non-existent-alias')).toBe(false)
+    })
+
+    it('处理批量注册时的部分失败', () => {
+      const mixedConfigs: ProviderConfig[] = [
+        {
+          id: 'valid-provider-1',
+          name: 'Valid Provider 1',
+          creator: vi.fn(() => ({ name: 'valid-1' })),
+          supportsImageGeneration: false
+        },
+        {
+          id: '', // 无效配置
+          name: 'Invalid Provider',
+          creator: vi.fn(() => ({ name: 'invalid' })),
+          supportsImageGeneration: false
+        } as any,
+        {
+          id: 'valid-provider-2',
+          name: 'Valid Provider 2',
+          creator: vi.fn(() => ({ name: 'valid-2' })),
+          supportsImageGeneration: true
+        }
+      ]
+
+      const successCount = registerMultipleProviderConfigs(mixedConfigs)
+      expect(successCount).toBe(2) // 只有两个有效配置成功
+
+      expect(hasProviderConfig('valid-provider-1')).toBe(true)
+      expect(hasProviderConfig('valid-provider-2')).toBe(true)
+      expect(hasProviderConfig('')).toBe(false)
+    })
+
+    it('处理动态导入失败的情况', async () => {
+      const config: ProviderConfig = {
+        id: 'import-test-provider',
+        name: 'Import Test Provider',
+        import: vi.fn().mockRejectedValue(new Error('Import failed')),
+        creatorFunctionName: 'createTest',
+        supportsImageGeneration: false
+      }
+
+      registerProviderConfig(config)
+
+      await expect(createProvider('import-test-provider', {})).rejects.toThrow('Import failed')
+    })
+  })
+
+  describe('集成测试', () => {
+    it('正确处理复杂的配置、创建、注册和清理场景', async () => {
+      // 初始状态验证
+      const initialConfigs = getAllProviderConfigs()
+      expect(initialConfigs.length).toBeGreaterThan(0) // 有基础配置
+      expect(getInitializedProviders()).toEqual([])
+
+      // 注册多个带别名的 provider 配置
+      const configs: ProviderConfig[] = [
+        {
+          id: 'integration-provider-1',
+          name: 'Integration Provider 1',
+          creator: vi.fn(() => ({ name: 'integration-1' })),
+          supportsImageGeneration: false,
+          aliases: ['alias-1', 'short-name-1']
+        },
+        {
+          id: 'integration-provider-2',
+          name: 'Integration Provider 2',
+          creator: vi.fn(() => ({ name: 'integration-2' })),
+          supportsImageGeneration: true,
+          aliases: ['alias-2', 'short-name-2']
+        }
+      ]
+
+      const successCount = registerMultipleProviderConfigs(configs)
+      expect(successCount).toBe(2)
+
+      // 验证配置注册成功
+      expect(hasProviderConfig('integration-provider-1')).toBe(true)
+      expect(hasProviderConfig('integration-provider-2')).toBe(true)
+      expect(hasProviderConfigByAlias('alias-1')).toBe(true)
+      expect(hasProviderConfigByAlias('alias-2')).toBe(true)
+
+      // 验证别名映射
+      const aliases = getAllProviderConfigAliases()
+      expect(aliases['alias-1']).toBe('integration-provider-1')
+      expect(aliases['alias-2']).toBe('integration-provider-2')
+
+      // 创建和注册 providers
+      const success1 = await createAndRegisterProvider('integration-provider-1', { apiKey: 'test1' })
+      const success2 = await createAndRegisterProvider('integration-provider-2', { apiKey: 'test2' })
+      expect(success1).toBe(true)
+      expect(success2).toBe(true)
+
+      // 验证注册成功
+      const registeredProviders = getInitializedProviders()
+      expect(registeredProviders).toContain('integration-provider-1')
+      expect(registeredProviders).toContain('integration-provider-2')
+      expect(hasInitializedProviders()).toBe(true)
+
+      // 清理
+      cleanup()
+
+      // 验证清理后的状态
+      expect(getInitializedProviders()).toEqual([])
+      expect(hasProviderConfig('integration-provider-1')).toBe(false)
+      expect(hasProviderConfig('integration-provider-2')).toBe(false)
+      expect(getAllProviderConfigAliases()).toEqual({})
+
+      // 基础配置应该重新加载
+      expect(hasProviderConfig('openai')).toBe(true)
+    })
+
+    it('正确处理动态导入配置的 provider', async () => {
+      const mockModule = { createCustomProvider: vi.fn(() => ({ name: 'custom-dynamic' })) }
+      const dynamicImportConfig: ProviderConfig = {
+        id: 'dynamic-import-provider',
+        name: 'Dynamic Import Provider',
+        import: vi.fn().mockResolvedValue(mockModule),
+        creatorFunctionName: 'createCustomProvider',
+        supportsImageGeneration: false
+      }
+
+      // 注册配置
+      const configSuccess = registerProviderConfig(dynamicImportConfig)
+      expect(configSuccess).toBe(true)
+
+      // 创建和注册 provider
+      const registerSuccess = await createAndRegisterProvider('dynamic-import-provider', { apiKey: 'test' })
+      expect(registerSuccess).toBe(true)
+
+      // 验证导入函数被调用
+      expect(dynamicImportConfig.import).toHaveBeenCalled()
+      expect(mockModule.createCustomProvider).toHaveBeenCalledWith({ apiKey: 'test' })
+
+      // 验证注册成功
+      expect(getInitializedProviders()).toContain('dynamic-import-provider')
+    })
+
+    it('正确处理大量配置的注册和管理', () => {
+      const largeConfigList: ProviderConfig[] = []
+
+      // 生成50个配置
+      for (let i = 0; i < 50; i++) {
+        largeConfigList.push({
+          id: `bulk-provider-${i}`,
+          name: `Bulk Provider ${i}`,
+          creator: vi.fn(() => ({ name: `bulk-${i}` })),
+          supportsImageGeneration: i % 2 === 0, // 偶数支持图像生成
+          aliases: [`alias-${i}`, `short-${i}`]
+        })
+      }
+
+      const successCount = registerMultipleProviderConfigs(largeConfigList)
+      expect(successCount).toBe(50)
+
+      // 验证所有配置都被正确注册
+      const allConfigs = getAllProviderConfigs()
+      expect(allConfigs.filter((config) => config.id.startsWith('bulk-provider-')).length).toBe(50)
+
+      // 验证别名数量
+      const aliases = getAllProviderConfigAliases()
+      const bulkAliases = Object.keys(aliases).filter(
+        (alias) => alias.startsWith('alias-') || alias.startsWith('short-')
+      )
+      expect(bulkAliases.length).toBe(100) // 每个 provider 有2个别名
+
+      // 随机验证几个配置
+      expect(hasProviderConfig('bulk-provider-0')).toBe(true)
+      expect(hasProviderConfig('bulk-provider-25')).toBe(true)
+      expect(hasProviderConfig('bulk-provider-49')).toBe(true)
+
+      // 验证别名工作正常
+      expect(resolveProviderConfigId('alias-25')).toBe('bulk-provider-25')
+      expect(isProviderConfigAlias('short-30')).toBe(true)
+
+      // 清理能正确处理大量数据
+      cleanup()
+      const cleanupAliases = getAllProviderConfigAliases()
+      expect(
+        Object.keys(cleanupAliases).filter((alias) => alias.startsWith('alias-') || alias.startsWith('short-'))
+      ).toEqual([])
+    })
+  })
+
+  describe('边界测试', () => {
+    it('处理包含特殊字符的 provider IDs', () => {
+      const specialCharsConfigs: ProviderConfig[] = [
+        {
+          id: 'provider-with-dashes',
+          name: 'Provider With Dashes',
+          creator: vi.fn(() => ({ name: 'dashes' })),
+          supportsImageGeneration: false
+        },
+        {
+          id: 'provider_with_underscores',
+          name: 'Provider With Underscores',
+          creator: vi.fn(() => ({ name: 'underscores' })),
+          supportsImageGeneration: false
+        },
+        {
+          id: 'provider.with.dots',
+          name: 'Provider With Dots',
+          creator: vi.fn(() => ({ name: 'dots' })),
+          supportsImageGeneration: false
+        }
+      ]
+
+      const successCount = registerMultipleProviderConfigs(specialCharsConfigs)
+      expect(successCount).toBeGreaterThan(0) // 至少有一些成功
+
+      // 验证支持的特殊字符格式
+      if (hasProviderConfig('provider-with-dashes')) {
+        expect(getProviderConfig('provider-with-dashes')).toBeDefined()
+      }
+      if (hasProviderConfig('provider_with_underscores')) {
+        expect(getProviderConfig('provider_with_underscores')).toBeDefined()
+      }
+    })
+
+    it('处理空的批量注册', () => {
+      const successCount = registerMultipleProviderConfigs([])
+      expect(successCount).toBe(0)
+
+      // 确保没有额外的配置被添加
+      const configsBefore = getAllProviderConfigs().length
+      expect(configsBefore).toBeGreaterThan(0) // 应该有基础配置
+    })
+
+    it('处理重复的配置注册', () => {
+      const config: ProviderConfig = {
+        id: 'duplicate-test-provider',
+        name: 'Duplicate Test Provider',
+        creator: vi.fn(() => ({ name: 'duplicate' })),
+        supportsImageGeneration: false
+      }
+
+      // 第一次注册成功
+      expect(registerProviderConfig(config)).toBe(true)
+      expect(hasProviderConfig('duplicate-test-provider')).toBe(true)
+
+      // 重复注册相同的配置（允许覆盖）
+      const updatedConfig: ProviderConfig = {
+        ...config,
+        name: 'Updated Duplicate Test Provider'
+      }
+      expect(registerProviderConfig(updatedConfig)).toBe(true)
+      expect(hasProviderConfig('duplicate-test-provider')).toBe(true)
+
+      // 验证配置被更新
+      const retrievedConfig = getProviderConfig('duplicate-test-provider')
+      expect(retrievedConfig?.name).toBe('Updated Duplicate Test Provider')
+    })
+
+    it('处理极长的 ID 和名称', () => {
+      const longId = 'very-long-provider-id-' + 'x'.repeat(100)
+      const longName = 'Very Long Provider Name ' + 'Y'.repeat(100)
+
+      const config: ProviderConfig = {
+        id: longId,
+        name: longName,
+        creator: vi.fn(() => ({ name: 'long-test' })),
+        supportsImageGeneration: false
+      }
+
+      const success = registerProviderConfig(config)
+      expect(success).toBe(true)
+      expect(hasProviderConfig(longId)).toBe(true)
+
+      const retrievedConfig = getProviderConfig(longId)
+      expect(retrievedConfig?.name).toBe(longName)
+    })
+
+    it('处理大量别名的配置', () => {
+      const manyAliases = Array.from({ length: 50 }, (_, i) => `alias-${i}`)
+
+      const config: ProviderConfig = {
+        id: 'provider-with-many-aliases',
+        name: 'Provider With Many Aliases',
+        creator: vi.fn(() => ({ name: 'many-aliases' })),
+        supportsImageGeneration: false,
+        aliases: manyAliases
+      }
+
+      const success = registerProviderConfig(config)
+      expect(success).toBe(true)
+
+      // 验证所有别名都能正确解析
+      manyAliases.forEach((alias) => {
+        expect(hasProviderConfigByAlias(alias)).toBe(true)
+        expect(resolveProviderConfigId(alias)).toBe('provider-with-many-aliases')
+        expect(isProviderConfigAlias(alias)).toBe(true)
+      })
+    })
+  })
+})

packages/aiCore/src/core/providers/__tests__/schemas.test.ts

@@ -0,0 +1,264 @@
+import { describe, expect, it, vi } from 'vitest'
+
+import {
+  type BaseProviderId,
+  baseProviderIds,
+  baseProviderIdSchema,
+  baseProviders,
+  type CustomProviderId,
+  customProviderIdSchema,
+  providerConfigSchema,
+  type ProviderId,
+  providerIdSchema
+} from '../schemas'
+
+describe('Provider Schemas', () => {
+  describe('baseProviders', () => {
+    it('包含所有预期的基础 providers', () => {
+      expect(baseProviders).toBeDefined()
+      expect(Array.isArray(baseProviders)).toBe(true)
+      expect(baseProviders.length).toBeGreaterThan(0)
+
+      const expectedIds = [
+        'openai',
+        'openai-responses',
+        'openai-compatible',
+        'anthropic',
+        'google',
+        'xai',
+        'azure',
+        'deepseek'
+      ]
+      const actualIds = baseProviders.map((p) => p.id)
+      expectedIds.forEach((id) => {
+        expect(actualIds).toContain(id)
+      })
+    })
+
+    it('每个基础 provider 有必要的属性', () => {
+      baseProviders.forEach((provider) => {
+        expect(provider).toHaveProperty('id')
+        expect(provider).toHaveProperty('name')
+        expect(provider).toHaveProperty('creator')
+        expect(provider).toHaveProperty('supportsImageGeneration')
+
+        expect(typeof provider.id).toBe('string')
+        expect(typeof provider.name).toBe('string')
+        expect(typeof provider.creator).toBe('function')
+        expect(typeof provider.supportsImageGeneration).toBe('boolean')
+      })
+    })
+
+    it('provider ID 是唯一的', () => {
+      const ids = baseProviders.map((p) => p.id)
+      const uniqueIds = [...new Set(ids)]
+      expect(ids).toEqual(uniqueIds)
+    })
+  })
+
+  describe('baseProviderIds', () => {
+    it('正确提取所有基础 provider IDs', () => {
+      expect(baseProviderIds).toBeDefined()
+      expect(Array.isArray(baseProviderIds)).toBe(true)
+      expect(baseProviderIds.length).toBe(baseProviders.length)
+
+      baseProviders.forEach((provider) => {
+        expect(baseProviderIds).toContain(provider.id)
+      })
+    })
+  })
+
+  describe('baseProviderIdSchema', () => {
+    it('验证有效的基础 provider IDs', () => {
+      baseProviderIds.forEach((id) => {
+        expect(baseProviderIdSchema.safeParse(id).success).toBe(true)
+      })
+    })
+
+    it('拒绝无效的基础 provider IDs', () => {
+      const invalidIds = ['invalid', 'not-exists', '']
+      invalidIds.forEach((id) => {
+        expect(baseProviderIdSchema.safeParse(id).success).toBe(false)
+      })
+    })
+  })
+
+  describe('customProviderIdSchema', () => {
+    it('接受有效的自定义 provider IDs', () => {
+      const validIds = ['custom-provider', 'my-ai-service', 'company-llm-v2']
+      validIds.forEach((id) => {
+        expect(customProviderIdSchema.safeParse(id).success).toBe(true)
+      })
+    })
+
+    it('拒绝与基础 provider IDs 冲突的 IDs', () => {
+      baseProviderIds.forEach((id) => {
+        expect(customProviderIdSchema.safeParse(id).success).toBe(false)
+      })
+    })
+
+    it('拒绝空字符串', () => {
+      expect(customProviderIdSchema.safeParse('').success).toBe(false)
+    })
+  })
+
+  describe('providerIdSchema', () => {
+    it('接受基础 provider IDs', () => {
+      baseProviderIds.forEach((id) => {
+        expect(providerIdSchema.safeParse(id).success).toBe(true)
+      })
+    })
+
+    it('接受有效的自定义 provider IDs', () => {
+      const validCustomIds = ['custom-provider', 'my-ai-service']
+      validCustomIds.forEach((id) => {
+        expect(providerIdSchema.safeParse(id).success).toBe(true)
+      })
+    })
+
+    it('拒绝无效的 IDs', () => {
+      const invalidIds = ['', undefined, null, 123]
+      invalidIds.forEach((id) => {
+        expect(providerIdSchema.safeParse(id).success).toBe(false)
+      })
+    })
+  })
+
+  describe('providerConfigSchema', () => {
+    it('验证带有 creator 的有效配置', () => {
+      const validConfig = {
+        id: 'custom-provider',
+        name: 'Custom Provider',
+        creator: vi.fn(),
+        supportsImageGeneration: true
+      }
+      expect(providerConfigSchema.safeParse(validConfig).success).toBe(true)
+    })
+
+    it('验证带有 import 配置的有效配置', () => {
+      const validConfig = {
+        id: 'custom-provider',
+        name: 'Custom Provider',
+        import: vi.fn(),
+        creatorFunctionName: 'createCustom',
+        supportsImageGeneration: false
+      }
+      expect(providerConfigSchema.safeParse(validConfig).success).toBe(true)
+    })
+
+    it('拒绝既没有 creator 也没有 import 配置的配置', () => {
+      const invalidConfig = {
+        id: 'invalid',
+        name: 'Invalid Provider',
+        supportsImageGeneration: false
+      }
+      expect(providerConfigSchema.safeParse(invalidConfig).success).toBe(false)
+    })
+
+    it('为 supportsImageGeneration 设置默认值', () => {
+      const config = {
+        id: 'test',
+        name: 'Test',
+        creator: vi.fn()
+      }
+      const result = providerConfigSchema.safeParse(config)
+      expect(result.success).toBe(true)
+      if (result.success) {
+        expect(result.data.supportsImageGeneration).toBe(false)
+      }
+    })
+
+    it('拒绝使用基础 provider ID 的配置', () => {
+      const invalidConfig = {
+        id: 'openai', // 基础 provider ID
+        name: 'Should Fail',
+        creator: vi.fn()
+      }
+      expect(providerConfigSchema.safeParse(invalidConfig).success).toBe(false)
+    })
+
+    it('拒绝缺少必需字段的配置', () => {
+      const invalidConfigs = [
+        { name: 'Missing ID', creator: vi.fn() },
+        { id: 'missing-name', creator: vi.fn() },
+        { id: '', name: 'Empty ID', creator: vi.fn() },
+        { id: 'valid-custom', name: '', creator: vi.fn() }
+      ]
+
+      invalidConfigs.forEach((config) => {
+        expect(providerConfigSchema.safeParse(config).success).toBe(false)
+      })
+    })
+  })
+
+  describe('Schema 验证功能', () => {
+    it('baseProviderIdSchema 正确验证基础 provider IDs', () => {
+      baseProviderIds.forEach((id) => {
+        expect(baseProviderIdSchema.safeParse(id).success).toBe(true)
+      })
+
+      expect(baseProviderIdSchema.safeParse('invalid-id').success).toBe(false)
+    })
+
+    it('customProviderIdSchema 正确验证自定义 provider IDs', () => {
+      const customIds = ['custom-provider', 'my-service', 'company-llm']
+      customIds.forEach((id) => {
+        expect(customProviderIdSchema.safeParse(id).success).toBe(true)
+      })
+
+      // 拒绝基础 provider IDs
+      baseProviderIds.forEach((id) => {
+        expect(customProviderIdSchema.safeParse(id).success).toBe(false)
+      })
+    })
+
+    it('providerIdSchema 接受基础和自定义 provider IDs', () => {
+      // 基础 IDs
+      baseProviderIds.forEach((id) => {
+        expect(providerIdSchema.safeParse(id).success).toBe(true)
+      })
+
+      // 自定义 IDs
+      const customIds = ['custom-provider', 'my-service']
+      customIds.forEach((id) => {
+        expect(providerIdSchema.safeParse(id).success).toBe(true)
+      })
+    })
+
+    it('providerConfigSchema 验证完整的 provider 配置', () => {
+      const validConfig = {
+        id: 'custom-provider',
+        name: 'Custom Provider',
+        creator: vi.fn(),
+        supportsImageGeneration: true
+      }
+      expect(providerConfigSchema.safeParse(validConfig).success).toBe(true)
+
+      const invalidConfig = {
+        id: 'openai', // 不允许基础 provider ID
+        name: 'OpenAI',
+        creator: vi.fn()
+      }
+      expect(providerConfigSchema.safeParse(invalidConfig).success).toBe(false)
+    })
+  })
+
+  describe('类型推导', () => {
+    it('BaseProviderId 类型正确', () => {
+      const id: BaseProviderId = 'openai'
+      expect(baseProviderIds).toContain(id)
+    })
+
+    it('CustomProviderId 类型是字符串', () => {
+      const id: CustomProviderId = 'custom-provider'
+      expect(typeof id).toBe('string')
+    })
+
+    it('ProviderId 类型支持基础和自定义 IDs', () => {
+      const baseId: ProviderId = 'openai'
+      const customId: ProviderId = 'custom-provider'
+      expect(typeof baseId).toBe('string')
+      expect(typeof customId).toBe('string')
+    })
+  })
+})

packages/aiCore/src/core/providers/factory.ts

@@ -0,0 +1,291 @@
+/**
+ * AI Provider 配置工厂
+ * 提供类型安全的 Provider 配置构建器
+ */
+
+import type { ProviderId, ProviderSettingsMap } from './types'
+
+/**
+ * 通用配置基础类型，包含所有 Provider 共有的属性
+ */
+export interface BaseProviderConfig {
+  apiKey?: string
+  baseURL?: string
+  timeout?: number
+  headers?: Record<string, string>
+  fetch?: typeof globalThis.fetch
+}
+
+/**
+ * 完整的配置类型，结合基础配置、AI SDK 配置和特定 Provider 配置
+ */
+type CompleteProviderConfig<T extends ProviderId> = BaseProviderConfig & Partial<ProviderSettingsMap[T]>
+
+type ConfigHandler<T extends ProviderId> = (
+  builder: ProviderConfigBuilder<T>,
+  provider: CompleteProviderConfig<T>
+) => void
+
+const configHandlers: {
+  [K in ProviderId]?: ConfigHandler<K>
+} = {
+  azure: (builder, provider) => {
+    const azureBuilder = builder as ProviderConfigBuilder<'azure'>
+    const azureProvider = provider as CompleteProviderConfig<'azure'>
+    azureBuilder.withAzureConfig({
+      apiVersion: azureProvider.apiVersion,
+      resourceName: azureProvider.resourceName
+    })
+  }
+}
+
+export class ProviderConfigBuilder<T extends ProviderId = ProviderId> {
+  private config: CompleteProviderConfig<T> = {} as CompleteProviderConfig<T>
+
+  constructor(private providerId: T) {}
+
+  /**
+   * 设置 API Key
+   */
+  withApiKey(apiKey: string): this
+  withApiKey(apiKey: string, options: T extends 'openai' ? { organization?: string; project?: string } : never): this
+  withApiKey(apiKey: string, options?: any): this {
+    this.config.apiKey = apiKey
+
+    // 类型安全的 OpenAI 特定配置
+    if (this.providerId === 'openai' && options) {
+      const openaiConfig = this.config as CompleteProviderConfig<'openai'>
+      if (options.organization) {
+        openaiConfig.organization = options.organization
+      }
+      if (options.project) {
+        openaiConfig.project = options.project
+      }
+    }
+
+    return this
+  }
+
+  /**
+   * 设置基础 URL
+   */
+  withBaseURL(baseURL: string) {
+    this.config.baseURL = baseURL
+    return this
+  }
+
+  /**
+   * 设置请求配置
+   */
+  withRequestConfig(options: { headers?: Record<string, string>; fetch?: typeof fetch }): this {
+    if (options.headers) {
+      this.config.headers = { ...this.config.headers, ...options.headers }
+    }
+    if (options.fetch) {
+      this.config.fetch = options.fetch
+    }
+    return this
+  }
+
+  /**
+   * Azure OpenAI 特定配置
+   */
+  withAzureConfig(options: { apiVersion?: string; resourceName?: string }): T extends 'azure' ? this : never
+  withAzureConfig(options: any): any {
+    if (this.providerId === 'azure') {
+      const azureConfig = this.config as CompleteProviderConfig<'azure'>
+      if (options.apiVersion) {
+        azureConfig.apiVersion = options.apiVersion
+      }
+      if (options.resourceName) {
+        azureConfig.resourceName = options.resourceName
+      }
+    }
+    return this
+  }
+
+  /**
+   * 设置自定义参数
+   */
+  withCustomParams(params: Record<string, any>) {
+    Object.assign(this.config, params)
+    return this
+  }
+
+  /**
+   * 构建最终配置
+   */
+  build(): ProviderSettingsMap[T] {
+    return this.config as ProviderSettingsMap[T]
+  }
+}
+
+/**
+ * Provider 配置工厂
+ * 提供便捷的配置创建方法
+ */
+export class ProviderConfigFactory {
+  /**
+   * 创建配置构建器
+   */
+  static builder<T extends ProviderId>(providerId: T): ProviderConfigBuilder<T> {
+    return new ProviderConfigBuilder(providerId)
+  }
+
+  /**
+   * 从通用Provider对象创建配置 - 使用更优雅的处理器模式
+   */
+  static fromProvider<T extends ProviderId>(
+    providerId: T,
+    provider: CompleteProviderConfig<T>,
+    options?: {
+      headers?: Record<string, string>
+      [key: string]: any
+    }
+  ): ProviderSettingsMap[T] {
+    const builder = new ProviderConfigBuilder<T>(providerId)
+
+    // 设置基本配置
+    if (provider.apiKey) {
+      builder.withApiKey(provider.apiKey)
+    }
+
+    if (provider.baseURL) {
+      builder.withBaseURL(provider.baseURL)
+    }
+
+    // 设置请求配置
+    if (options?.headers) {
+      builder.withRequestConfig({
+        headers: options.headers
+      })
+    }
+
+    // 使用配置处理器模式 - 更加优雅和可扩展
+    const handler = configHandlers[providerId]
+    if (handler) {
+      handler(builder, provider)
+    }
+
+    // 添加其他自定义参数
+    if (options) {
+      const customOptions = { ...options }
+      delete customOptions.headers // 已经处理过了
+      if (Object.keys(customOptions).length > 0) {
+        builder.withCustomParams(customOptions)
+      }
+    }
+
+    return builder.build()
+  }
+
+  /**
+   * 快速创建 OpenAI 配置
+   */
+  static createOpenAI(
+    apiKey: string,
+    options?: {
+      baseURL?: string
+      organization?: string
+      project?: string
+    }
+  ) {
+    const builder = this.builder('openai')
+
+    // 使用类型安全的重载
+    if (options?.organization || options?.project) {
+      builder.withApiKey(apiKey, {
+        organization: options.organization,
+        project: options.project
+      })
+    } else {
+      builder.withApiKey(apiKey)
+    }
+
+    return builder.withBaseURL(options?.baseURL || 'https://api.openai.com').build()
+  }
+
+  /**
+   * 快速创建 Anthropic 配置
+   */
+  static createAnthropic(
+    apiKey: string,
+    options?: {
+      baseURL?: string
+    }
+  ) {
+    return this.builder('anthropic')
+      .withApiKey(apiKey)
+      .withBaseURL(options?.baseURL || 'https://api.anthropic.com')
+      .build()
+  }
+
+  /**
+   * 快速创建 Azure OpenAI 配置
+   */
+  static createAzureOpenAI(
+    apiKey: string,
+    options: {
+      baseURL: string
+      apiVersion?: string
+      resourceName?: string
+    }
+  ) {
+    return this.builder('azure')
+      .withApiKey(apiKey)
+      .withBaseURL(options.baseURL)
+      .withAzureConfig({
+        apiVersion: options.apiVersion,
+        resourceName: options.resourceName
+      })
+      .build()
+  }
+
+  /**
+   * 快速创建 Google 配置
+   */
+  static createGoogle(
+    apiKey: string,
+    options?: {
+      baseURL?: string
+      projectId?: string
+      location?: string
+    }
+  ) {
+    return this.builder('google')
+      .withApiKey(apiKey)
+      .withBaseURL(options?.baseURL || 'https://generativelanguage.googleapis.com')
+      .build()
+  }
+
+  /**
+   * 快速创建 Vertex AI 配置
+   */
+  static createVertexAI() {
+    // credentials: {
+    //   clientEmail: string
+    //   privateKey: string
+    // },
+    // options?: {
+    //   project?: string
+    //   location?: string
+    // }
+    // return this.builder('google-vertex')
+    //   .withGoogleCredentials(credentials)
+    //   .withGoogleVertexConfig({
+    //     project: options?.project,
+    //     location: options?.location
+    //   })
+    //   .build()
+  }
+
+  static createOpenAICompatible(baseURL: string, apiKey: string) {
+    return this.builder('openai-compatible').withBaseURL(baseURL).withApiKey(apiKey).build()
+  }
+}
+
+/**
+ * 便捷的配置创建函数
+ */
+export const createProviderConfig = ProviderConfigFactory.fromProvider
+export const providerConfigBuilder = ProviderConfigFactory.builder

packages/aiCore/src/core/providers/index.ts

@@ -0,0 +1,83 @@
+/**
+ * Providers 模块统一导出 - 独立Provider包
+ */
+
+// ==================== 核心管理器 ====================
+
+// Provider 注册表管理器
+export { globalRegistryManagement, RegistryManagement } from './RegistryManagement'
+
+// Provider 核心功能
+export {
+  // 状态管理
+  cleanup,
+  clearAllProviders,
+  createAndRegisterProvider,
+  createProvider,
+  getAllProviderConfigAliases,
+  getAllProviderConfigs,
+  getImageModel,
+  // 工具函数
+  getInitializedProviders,
+  getLanguageModel,
+  getProviderConfig,
+  getProviderConfigByAlias,
+  getSupportedProviders,
+  getTextEmbeddingModel,
+  hasInitializedProviders,
+  // 工具函数
+  hasProviderConfig,
+  // 别名支持
+  hasProviderConfigByAlias,
+  isProviderConfigAlias,
+  // 错误类型
+  ProviderInitializationError,
+  // 全局访问
+  providerRegistry,
+  registerMultipleProviderConfigs,
+  registerProvider,
+  // 统一Provider系统
+  registerProviderConfig,
+  resolveProviderConfigId
+} from './registry'
+
+// ==================== 基础数据和类型 ====================
+
+// 基础Provider数据源
+export { baseProviderIds, baseProviders } from './schemas'
+
+// 类型定义和Schema
+export type {
+  BaseProviderId,
+  CustomProviderId,
+  DynamicProviderRegistration,
+  ProviderConfig,
+  ProviderId
+} from './schemas' // 从 schemas 导出的类型
+export { baseProviderIdSchema, customProviderIdSchema, providerConfigSchema, providerIdSchema } from './schemas' // Schema 导出
+export type {
+  DynamicProviderRegistry,
+  ExtensibleProviderSettingsMap,
+  ProviderError,
+  ProviderSettingsMap,
+  ProviderTypeRegistrar
+} from './types'
+
+// ==================== 工具函数 ====================
+
+// Provider配置工厂
+export {
+  type BaseProviderConfig,
+  createProviderConfig,
+  ProviderConfigBuilder,
+  providerConfigBuilder,
+  ProviderConfigFactory
+} from './factory'
+
+// 工具函数
+export { formatPrivateKey } from './utils'
+
+// ==================== 扩展功能 ====================
+
+// Hub Provider 功能
+export { createHubProvider, type HubProviderConfig, HubProviderError } from './HubProvider'

packages/aiCore/src/core/providers/registry.ts

@@ -0,0 +1,320 @@
+/**
+ * Provider 初始化器
+ * 负责根据配置创建 providers 并注册到全局管理器
+ * 集成了来自 ModelCreator 的特殊处理逻辑
+ */
+
+import { customProvider } from 'ai'
+
+import { globalRegistryManagement } from './RegistryManagement'
+import { baseProviders, type ProviderConfig } from './schemas'
+
+/**
+ * Provider 初始化错误类型
+ */
+class ProviderInitializationError extends Error {
+  constructor(
+    message: string,
+    public providerId?: string,
+    public cause?: Error
+  ) {
+    super(message)
+    this.name = 'ProviderInitializationError'
+  }
+}
+
+// ==================== 全局管理器导出 ====================
+
+export { globalRegistryManagement as providerRegistry }
+
+// ==================== 便捷访问方法 ====================
+
+export const getLanguageModel = (id: string) => globalRegistryManagement.languageModel(id as any)
+export const getTextEmbeddingModel = (id: string) => globalRegistryManagement.textEmbeddingModel(id as any)
+export const getImageModel = (id: string) => globalRegistryManagement.imageModel(id as any)
+
+// ==================== 工具函数 ====================
+
+/**
+ * 获取支持的 Providers 列表
+ */
+export function getSupportedProviders(): Array<{
+  id: string
+  name: string
+}> {
+  return baseProviders.map((provider) => ({
+    id: provider.id,
+    name: provider.name
+  }))
+}
+
+/**
+ * 获取所有已初始化的 providers
+ */
+export function getInitializedProviders(): string[] {
+  return globalRegistryManagement.getRegisteredProviders()
+}
+
+/**
+ * 检查是否有任何已初始化的 providers
+ */
+export function hasInitializedProviders(): boolean {
+  return globalRegistryManagement.hasProviders()
+}
+
+// ==================== 统一Provider配置系统 ====================
+
+// 全局Provider配置存储
+const providerConfigs = new Map<string, ProviderConfig>()
+// 全局ProviderConfig别名映射 - 借鉴RegistryManagement模式
+const providerConfigAliases = new Map<string, string>() // alias -> realId
+
+/**
+ * 初始化内置配置 - 将baseProviders转换为统一格式
+ */
+function initializeBuiltInConfigs(): void {
+  baseProviders.forEach((provider) => {
+    const config: ProviderConfig = {
+      id: provider.id,
+      name: provider.name,
+      creator: provider.creator as any, // 类型转换以兼容多种creator签名
+      supportsImageGeneration: provider.supportsImageGeneration || false
+    }
+    providerConfigs.set(provider.id, config)
+  })
+}
+
+// 启动时自动注册内置配置
+initializeBuiltInConfigs()
+
+/**
+ * 步骤1: 注册Provider配置 - 仅存储配置，不执行创建
+ */
+export function registerProviderConfig(config: ProviderConfig): boolean {
+  try {
+    // 验证配置
+    if (!config || !config.id || !config.name) {
+      return false
+    }
+
+    // 检查是否与已有配置冲突（包括内置配置）
+    if (providerConfigs.has(config.id)) {
+      console.warn(`ProviderConfig "${config.id}" already exists, will override`)
+    }
+
+    // 存储配置（内置和用户配置统一处理）
+    providerConfigs.set(config.id, config)
+
+    //  处理别名
+    if (config.aliases && config.aliases.length > 0) {
+      config.aliases.forEach((alias) => {
+        if (providerConfigAliases.has(alias)) {
+          console.warn(`ProviderConfig alias "${alias}" already exists, will override`)
+        }
+        providerConfigAliases.set(alias, config.id)
+      })
+    }
+
+    return true
+  } catch (error) {
+    console.error(`Failed to register ProviderConfig:`, error)
+    return false
+  }
+}
+
+/**
+ * 步骤2: 创建Provider - 根据配置执行实际创建
+ */
+export async function createProvider(providerId: string, options: any): Promise<any> {
+  //  支持通过别名查找配置
+  const config = getProviderConfigByAlias(providerId)
+
+  if (!config) {
+    throw new Error(`ProviderConfig not found for id: ${providerId}`)
+  }
+
+  try {
+    let creator: (options: any) => any
+
+    if (config.creator) {
+      // 方式1: 直接执行 creator
+      creator = config.creator
+    } else if (config.import && config.creatorFunctionName) {
+      // 方式2: 动态导入并执行
+      const module = await config.import()
+      creator = (module as any)[config.creatorFunctionName]
+
+      if (!creator || typeof creator !== 'function') {
+        throw new Error(`Creator function "${config.creatorFunctionName}" not found in imported module`)
+      }
+    } else {
+      throw new Error('No valid creator method provided in ProviderConfig')
+    }
+
+    // 使用真实配置创建provider实例
+    return creator(options)
+  } catch (error) {
+    console.error(`Failed to create provider "${providerId}":`, error)
+    throw error
+  }
+}
+
+/**
+ * 步骤3: 注册Provider到全局管理器
+ */
+export function registerProvider(providerId: string, provider: any): boolean {
+  try {
+    const config = providerConfigs.get(providerId)
+    if (!config) {
+      console.error(`ProviderConfig not found for id: ${providerId}`)
+      return false
+    }
+
+    // 获取aliases配置
+    const aliases = config.aliases
+
+    // 处理特殊provider逻辑
+    if (providerId === 'openai') {
+      // 注册默认 openai
+      globalRegistryManagement.registerProvider(providerId, provider, aliases)
+
+      // 创建并注册 openai-chat 变体
+      const openaiChatProvider = customProvider({
+        fallbackProvider: {
+          ...provider,
+          languageModel: (modelId: string) => provider.chat(modelId)
+        }
+      })
+      globalRegistryManagement.registerProvider(`${providerId}-chat`, openaiChatProvider)
+    } else if (providerId === 'azure') {
+      globalRegistryManagement.registerProvider(`${providerId}-chat`, provider, aliases)
+      // 跟上面相反,creator产出的默认会调用chat
+      const azureResponsesProvider = customProvider({
+        fallbackProvider: {
+          ...provider,
+          languageModel: (modelId: string) => provider.responses(modelId)
+        }
+      })
+      globalRegistryManagement.registerProvider(providerId, azureResponsesProvider)
+    } else {
+      // 其他provider直接注册
+      globalRegistryManagement.registerProvider(providerId, provider, aliases)
+    }
+
+    return true
+  } catch (error) {
+    console.error(`Failed to register provider "${providerId}" to global registry:`, error)
+    return false
+  }
+}
+
+/**
+ * 便捷函数: 一次性完成创建+注册
+ */
+export async function createAndRegisterProvider(providerId: string, options: any): Promise<boolean> {
+  try {
+    // 步骤2: 创建provider
+    const provider = await createProvider(providerId, options)
+
+    // 步骤3: 注册到全局管理器
+    return registerProvider(providerId, provider)
+  } catch (error) {
+    console.error(`Failed to create and register provider "${providerId}":`, error)
+    return false
+  }
+}
+
+/**
+ * 批量注册Provider配置
+ */
+export function registerMultipleProviderConfigs(configs: ProviderConfig[]): number {
+  let successCount = 0
+  configs.forEach((config) => {
+    if (registerProviderConfig(config)) {
+      successCount++
+    }
+  })
+  return successCount
+}
+
+/**
+ * 检查是否有对应的Provider配置
+ */
+export function hasProviderConfig(providerId: string): boolean {
+  return providerConfigs.has(providerId)
+}
+
+/**
+ * 通过别名或ID检查是否有对应的Provider配置
+ */
+export function hasProviderConfigByAlias(aliasOrId: string): boolean {
+  const realId = resolveProviderConfigId(aliasOrId)
+  return providerConfigs.has(realId)
+}
+
+/**
+ * 获取所有Provider配置
+ */
+export function getAllProviderConfigs(): ProviderConfig[] {
+  return Array.from(providerConfigs.values())
+}
+
+/**
+ * 根据ID获取Provider配置
+ */
+export function getProviderConfig(providerId: string): ProviderConfig | undefined {
+  return providerConfigs.get(providerId)
+}
+
+/**
+ * 通过别名或ID获取Provider配置
+ */
+export function getProviderConfigByAlias(aliasOrId: string): ProviderConfig | undefined {
+  // 先检查是否为别名，如果是则解析为真实ID
+  const realId = providerConfigAliases.get(aliasOrId) || aliasOrId
+  return providerConfigs.get(realId)
+}
+
+/**
+ * 解析真实的ProviderConfig ID（去别名化）
+ */
+export function resolveProviderConfigId(aliasOrId: string): string {
+  return providerConfigAliases.get(aliasOrId) || aliasOrId
+}
+
+/**
+ * 检查是否为ProviderConfig别名
+ */
+export function isProviderConfigAlias(id: string): boolean {
+  return providerConfigAliases.has(id)
+}
+
+/**
+ * 获取所有ProviderConfig别名映射关系
+ */
+export function getAllProviderConfigAliases(): Record<string, string> {
+  const result: Record<string, string> = {}
+  providerConfigAliases.forEach((realId, alias) => {
+    result[alias] = realId
+  })
+  return result
+}
+
+/**
+ * 清理所有Provider配置和已注册的providers
+ */
+export function cleanup(): void {
+  providerConfigs.clear()
+  providerConfigAliases.clear() //  清理别名映射
+  globalRegistryManagement.clear()
+  // 重新初始化内置配置
+  initializeBuiltInConfigs()
+}
+
+export function clearAllProviders(): void {
+  globalRegistryManagement.clear()
+}
+
+// ==================== 导出错误类型 ====================
+
+export { ProviderInitializationError }

packages/aiCore/src/core/providers/schemas.ts

@@ -0,0 +1,194 @@
+/**
+ * Provider Config 定义
+ */
+
+import { createAnthropic } from '@ai-sdk/anthropic'
+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 { createOpenAI, type OpenAIProviderSettings } from '@ai-sdk/openai'
+import { createOpenAICompatible } from '@ai-sdk/openai-compatible'
+import { LanguageModelV2 } from '@ai-sdk/provider'
+import { createXai } from '@ai-sdk/xai'
+import { createOpenRouter } from '@openrouter/ai-sdk-provider'
+import { customProvider, Provider } from 'ai'
+import { z } from 'zod'
+
+/**
+ * 基础 Provider IDs
+ */
+export const baseProviderIds = [
+  'openai',
+  'openai-chat',
+  'openai-compatible',
+  'anthropic',
+  'google',
+  'xai',
+  'azure',
+  'azure-responses',
+  'deepseek',
+  'openrouter'
+] as const
+
+/**
+ * 基础 Provider ID Schema
+ */
+export const baseProviderIdSchema = z.enum(baseProviderIds)
+
+/**
+ * 基础 Provider ID
+ */
+export type BaseProviderId = z.infer<typeof baseProviderIdSchema>
+
+export const isBaseProvider = (id: ProviderId): id is BaseProviderId => {
+  return baseProviderIdSchema.safeParse(id).success
+}
+
+type BaseProvider = {
+  id: BaseProviderId
+  name: string
+  creator: (options: any) => Provider | LanguageModelV2
+  supportsImageGeneration: boolean
+}
+
+/**
+ * 基础 Providers 定义
+ * 作为唯一数据源，避免重复维护
+ */
+export const baseProviders = [
+  {
+    id: 'openai',
+    name: 'OpenAI',
+    creator: createOpenAI,
+    supportsImageGeneration: true
+  },
+  {
+    id: 'openai-chat',
+    name: 'OpenAI Chat',
+    creator: (options: OpenAIProviderSettings) => {
+      const provider = createOpenAI(options)
+      return customProvider({
+        fallbackProvider: {
+          ...provider,
+          languageModel: (modelId: string) => provider.chat(modelId)
+        }
+      })
+    },
+    supportsImageGeneration: true
+  },
+  {
+    id: 'openai-compatible',
+    name: 'OpenAI Compatible',
+    creator: createOpenAICompatible,
+    supportsImageGeneration: true
+  },
+  {
+    id: 'anthropic',
+    name: 'Anthropic',
+    creator: createAnthropic,
+    supportsImageGeneration: false
+  },
+  {
+    id: 'google',
+    name: 'Google Generative AI',
+    creator: createGoogleGenerativeAI,
+    supportsImageGeneration: true
+  },
+  {
+    id: 'xai',
+    name: 'xAI (Grok)',
+    creator: createXai,
+    supportsImageGeneration: true
+  },
+  {
+    id: 'azure',
+    name: 'Azure OpenAI',
+    creator: createAzure,
+    supportsImageGeneration: true
+  },
+  {
+    id: 'azure-responses',
+    name: 'Azure OpenAI Responses',
+    creator: (options: AzureOpenAIProviderSettings) => {
+      const provider = createAzure(options)
+      return customProvider({
+        fallbackProvider: {
+          ...provider,
+          languageModel: (modelId: string) => provider.responses(modelId)
+        }
+      })
+    },
+    supportsImageGeneration: true
+  },
+  {
+    id: 'deepseek',
+    name: 'DeepSeek',
+    creator: createDeepSeek,
+    supportsImageGeneration: false
+  },
+  {
+    id: 'openrouter',
+    name: 'OpenRouter',
+    creator: createOpenRouter,
+    supportsImageGeneration: true
+  }
+] as const satisfies BaseProvider[]
+
+/**
+ * 用户自定义 Provider ID Schema
+ * 允许任意字符串，但排除基础 provider IDs 以避免冲突
+ */
+export const customProviderIdSchema = z
+  .string()
+  .min(1)
+  .refine((id) => !baseProviderIds.includes(id as any), {
+    message: 'Custom provider ID cannot conflict with base provider IDs'
+  })
+
+/**
+ * Provider ID Schema - 支持基础和自定义
+ */
+export const providerIdSchema = z.union([baseProviderIdSchema, customProviderIdSchema])
+
+/**
+ * Provider 配置 Schema
+ * 用于Provider的配置验证
+ */
+export const providerConfigSchema = z
+  .object({
+    id: customProviderIdSchema, // 只允许自定义ID
+    name: z.string().min(1),
+    creator: z
+      .function({
+        input: z.any(),
+        output: z.any()
+      })
+      .optional(),
+    import: z.function().optional(),
+    creatorFunctionName: z.string().optional(),
+    supportsImageGeneration: z.boolean().default(false),
+    imageCreator: z.function().optional(),
+    validateOptions: z.function().optional(),
+    aliases: z.array(z.string()).optional()
+  })
+  .refine((data) => data.creator || (data.import && data.creatorFunctionName), {
+    message: 'Must provide either creator function or import configuration'
+  })
+
+/**
+ * Provider ID 类型 - 基于 zod schema 推导
+ */
+export type ProviderId = z.infer<typeof providerIdSchema>
+export type CustomProviderId = z.infer<typeof customProviderIdSchema>
+
+/**
+ * Provider 配置类型
+ */
+export type ProviderConfig = z.infer<typeof providerConfigSchema>
+
+/**
+ * 兼容性类型别名
+ * @deprecated 使用 ProviderConfig 替代
+ */
+export type DynamicProviderRegistration = ProviderConfig

packages/aiCore/src/core/providers/types.ts

@@ -0,0 +1,96 @@
+import { type AnthropicProviderSettings } from '@ai-sdk/anthropic'
+import { type AzureOpenAIProviderSettings } from '@ai-sdk/azure'
+import { type DeepSeekProviderSettings } from '@ai-sdk/deepseek'
+import { type GoogleGenerativeAIProviderSettings } from '@ai-sdk/google'
+import { type OpenAIProviderSettings } from '@ai-sdk/openai'
+import { type OpenAICompatibleProviderSettings } from '@ai-sdk/openai-compatible'
+import {
+  EmbeddingModelV2 as EmbeddingModel,
+  ImageModelV2 as ImageModel,
+  LanguageModelV2 as LanguageModel,
+  ProviderV2,
+  SpeechModelV2 as SpeechModel,
+  TranscriptionModelV2 as TranscriptionModel
+} from '@ai-sdk/provider'
+import { type XaiProviderSettings } from '@ai-sdk/xai'
+
+// 导入基于 Zod 的 ProviderId 类型
+import { type ProviderId as ZodProviderId } from './schemas'
+
+export interface ExtensibleProviderSettingsMap {
+  // 基础的静态providers
+  openai: OpenAIProviderSettings
+  'openai-responses': OpenAIProviderSettings
+  'openai-compatible': OpenAICompatibleProviderSettings
+  anthropic: AnthropicProviderSettings
+  google: GoogleGenerativeAIProviderSettings
+  xai: XaiProviderSettings
+  azure: AzureOpenAIProviderSettings
+  deepseek: DeepSeekProviderSettings
+}
+
+// 动态扩展的provider类型注册表
+export interface DynamicProviderRegistry {
+  [key: string]: any
+}
+
+// 合并基础和动态provider类型
+export type ProviderSettingsMap = ExtensibleProviderSettingsMap & DynamicProviderRegistry
+
+// 错误类型
+export class ProviderError extends Error {
+  constructor(
+    message: string,
+    public providerId: string,
+    public code?: string,
+    public cause?: Error
+  ) {
+    super(message)
+    this.name = 'ProviderError'
+  }
+}
+
+// 动态ProviderId类型 - 基于 Zod Schema，支持运行时扩展和验证
+export type ProviderId = ZodProviderId
+
+export interface ProviderTypeRegistrar {
+  registerProviderType<T extends string, S>(providerId: T, settingsType: S): void
+  getProviderSettings<T extends string>(providerId: T): any
+}
+
+// 重新导出所有类型供外部使用
+export type {
+  AnthropicProviderSettings,
+  AzureOpenAIProviderSettings,
+  DeepSeekProviderSettings,
+  GoogleGenerativeAIProviderSettings,
+  OpenAICompatibleProviderSettings,
+  OpenAIProviderSettings,
+  XaiProviderSettings
+}
+
+export type AiSdkModel = LanguageModel | ImageModel | EmbeddingModel<string> | TranscriptionModel | SpeechModel
+
+export type AiSdkModelType = 'text' | 'image' | 'embedding' | 'transcription' | 'speech'
+
+export const METHOD_MAP = {
+  text: 'languageModel',
+  image: 'imageModel',
+  embedding: 'textEmbeddingModel',
+  transcription: 'transcriptionModel',
+  speech: 'speechModel'
+} as const satisfies Record<AiSdkModelType, keyof ProviderV2>
+
+export type AiSdkModelMethodMap = Record<AiSdkModelType, keyof ProviderV2>
+
+export type AiSdkModelReturnMap = {
+  text: LanguageModel
+  image: ImageModel
+  embedding: EmbeddingModel<string>
+  transcription: TranscriptionModel
+  speech: SpeechModel
+}
+
+export type AiSdkMethodName<T extends AiSdkModelType> = (typeof METHOD_MAP)[T]
+
+export type AiSdkModelReturn<T extends AiSdkModelType> = AiSdkModelReturnMap[T]

packages/aiCore/src/core/providers/utils.ts

@@ -0,0 +1,86 @@
+/**
+ * 格式化私钥，确保它包含正确的PEM头部和尾部
+ */
+export function formatPrivateKey(privateKey: string): string {
+  if (!privateKey || typeof privateKey !== 'string') {
+    throw new Error('Private key must be a non-empty string')
+  }
+
+  // 先处理 JSON 字符串中的转义换行符
+  const key = privateKey.replace(/\\n/g, '\n')
+
+  // 检查是否已经是正确格式的 PEM 私钥
+  const hasBeginMarker = key.includes('-----BEGIN PRIVATE KEY-----')
+  const hasEndMarker = key.includes('-----END PRIVATE KEY-----')
+
+  if (hasBeginMarker && hasEndMarker) {
+    // 已经是 PEM 格式，但可能格式不规范，重新格式化
+    return normalizePemFormat(key)
+  }
+
+  // 如果没有完整的 PEM 头尾，尝试重新构建
+  return reconstructPemKey(key)
+}
+
+/**
+ * 标准化 PEM 格式
+ */
+function normalizePemFormat(pemKey: string): string {
+  // 分离头部、内容和尾部
+  const lines = pemKey
+    .split('\n')
+    .map((line) => line.trim())
+    .filter((line) => line.length > 0)
+
+  let keyContent = ''
+  let foundBegin = false
+  let foundEnd = false
+
+  for (const line of lines) {
+    if (line === '-----BEGIN PRIVATE KEY-----') {
+      foundBegin = true
+      continue
+    }
+    if (line === '-----END PRIVATE KEY-----') {
+      foundEnd = true
+      break
+    }
+    if (foundBegin && !foundEnd) {
+      keyContent += line
+    }
+  }
+
+  if (!foundBegin || !foundEnd || !keyContent) {
+    throw new Error('Invalid PEM format: missing BEGIN/END markers or key content')
+  }
+
+  // 重新格式化为 64 字符一行
+  const formattedContent = keyContent.match(/.{1,64}/g)?.join('\n') || keyContent
+
+  return `-----BEGIN PRIVATE KEY-----\n${formattedContent}\n-----END PRIVATE KEY-----`
+}
+
+/**
+ * 重新构建 PEM 私钥
+ */
+function reconstructPemKey(key: string): string {
+  // 移除所有空白字符和可能存在的不完整头尾
+  let cleanKey = key.replace(/\s+/g, '')
+  cleanKey = cleanKey.replace(/-----BEGIN[^-]*-----/g, '')
+  cleanKey = cleanKey.replace(/-----END[^-]*-----/g, '')
+
+  // 确保私钥内容不为空
+  if (!cleanKey) {
+    throw new Error('Private key content is empty after cleaning')
+  }
+
+  // 验证是否是有效的 Base64 字符
+  if (!/^[A-Za-z0-9+/=]+$/.test(cleanKey)) {
+    throw new Error('Private key contains invalid characters (not valid Base64)')
+  }
+
+  // 格式化为 64 字符一行
+  const formattedKey = cleanKey.match(/.{1,64}/g)?.join('\n') || cleanKey
+
+  return `-----BEGIN PRIVATE KEY-----\n${formattedKey}\n-----END PRIVATE KEY-----`
+}

packages/aiCore/src/core/runtime/__tests__/generateImage.test.ts

@@ -0,0 +1,523 @@
+import { ImageModelV2 } from '@ai-sdk/provider'
+import { experimental_generateImage as aiGenerateImage, NoImageGeneratedError } from 'ai'
+import { beforeEach, describe, expect, it, vi } from 'vitest'
+
+import { type AiPlugin } from '../../plugins'
+import { globalRegistryManagement } from '../../providers/RegistryManagement'
+import { ImageGenerationError, ImageModelResolutionError } from '../errors'
+import { RuntimeExecutor } from '../executor'
+
+// Mock dependencies
+vi.mock('ai', () => ({
+  experimental_generateImage: vi.fn(),
+  NoImageGeneratedError: class NoImageGeneratedError extends Error {
+    static isInstance = vi.fn()
+    constructor() {
+      super('No image generated')
+      this.name = 'NoImageGeneratedError'
+    }
+  }
+}))
+
+vi.mock('../../providers/RegistryManagement', () => ({
+  globalRegistryManagement: {
+    imageModel: vi.fn()
+  },
+  DEFAULT_SEPARATOR: '|'
+}))
+
+describe('RuntimeExecutor.generateImage', () => {
+  let executor: RuntimeExecutor<'openai'>
+  let mockImageModel: ImageModelV2
+  let mockGenerateImageResult: any
+
+  beforeEach(() => {
+    // Reset all mocks
+    vi.clearAllMocks()
+
+    // Create executor instance
+    executor = RuntimeExecutor.create('openai', {
+      apiKey: 'test-key'
+    })
+
+    // Mock image model
+    mockImageModel = {
+      modelId: 'dall-e-3',
+      provider: 'openai'
+    } as ImageModelV2
+
+    // Mock generateImage result
+    mockGenerateImageResult = {
+      image: {
+        base64: 'base64-encoded-image-data',
+        uint8Array: new Uint8Array([1, 2, 3]),
+        mediaType: 'image/png'
+      },
+      images: [
+        {
+          base64: 'base64-encoded-image-data',
+          uint8Array: new Uint8Array([1, 2, 3]),
+          mediaType: 'image/png'
+        }
+      ],
+      warnings: [],
+      providerMetadata: {
+        openai: {
+          images: [{ revisedPrompt: 'A detailed prompt' }]
+        }
+      },
+      responses: []
+    }
+
+    // Setup mocks to avoid "No providers registered" error
+    vi.mocked(globalRegistryManagement.imageModel).mockReturnValue(mockImageModel)
+    vi.mocked(aiGenerateImage).mockResolvedValue(mockGenerateImageResult)
+  })
+
+  describe('Basic functionality', () => {
+    it('should generate a single image with minimal parameters', async () => {
+      const result = await executor.generateImage({ model: 'dall-e-3', prompt: 'A futuristic cityscape at sunset' })
+
+      expect(globalRegistryManagement.imageModel).toHaveBeenCalledWith('openai|dall-e-3')
+
+      expect(aiGenerateImage).toHaveBeenCalledWith({
+        model: mockImageModel,
+        prompt: 'A futuristic cityscape at sunset'
+      })
+
+      expect(result).toEqual(mockGenerateImageResult)
+    })
+
+    it('should generate image with pre-created model', async () => {
+      const result = await executor.generateImage({
+        model: mockImageModel,
+        prompt: 'A beautiful landscape'
+      })
+
+      // Note: globalRegistryManagement.imageModel may still be called due to resolveImageModel logic
+      expect(aiGenerateImage).toHaveBeenCalledWith({
+        model: mockImageModel,
+        prompt: 'A beautiful landscape'
+      })
+
+      expect(result).toEqual(mockGenerateImageResult)
+    })
+
+    it('should support multiple images generation', async () => {
+      await executor.generateImage({ model: 'dall-e-3', prompt: 'A futuristic cityscape', n: 3 })
+
+      expect(aiGenerateImage).toHaveBeenCalledWith({
+        model: mockImageModel,
+        prompt: 'A futuristic cityscape',
+        n: 3
+      })
+    })
+
+    it('should support size specification', async () => {
+      await executor.generateImage({ model: 'dall-e-3', prompt: 'A beautiful sunset', size: '1024x1024' })
+
+      expect(aiGenerateImage).toHaveBeenCalledWith({
+        model: mockImageModel,
+        prompt: 'A beautiful sunset',
+        size: '1024x1024'
+      })
+    })
+
+    it('should support aspect ratio specification', async () => {
+      await executor.generateImage({ model: 'dall-e-3', prompt: 'A mountain landscape', aspectRatio: '16:9' })
+
+      expect(aiGenerateImage).toHaveBeenCalledWith({
+        model: mockImageModel,
+        prompt: 'A mountain landscape',
+        aspectRatio: '16:9'
+      })
+    })
+
+    it('should support seed for consistent output', async () => {
+      await executor.generateImage({ model: 'dall-e-3', prompt: 'A cat in space', seed: 1234567890 })
+
+      expect(aiGenerateImage).toHaveBeenCalledWith({
+        model: mockImageModel,
+        prompt: 'A cat in space',
+        seed: 1234567890
+      })
+    })
+
+    it('should support abort signal', async () => {
+      const abortController = new AbortController()
+
+      await executor.generateImage({ model: 'dall-e-3', prompt: 'A cityscape', abortSignal: abortController.signal })
+
+      expect(aiGenerateImage).toHaveBeenCalledWith({
+        model: mockImageModel,
+        prompt: 'A cityscape',
+        abortSignal: abortController.signal
+      })
+    })
+
+    it('should support provider-specific options', async () => {
+      await executor.generateImage({
+        model: 'dall-e-3',
+        prompt: 'A space station',
+        providerOptions: {
+          openai: {
+            quality: 'hd',
+            style: 'vivid'
+          }
+        }
+      })
+
+      expect(aiGenerateImage).toHaveBeenCalledWith({
+        model: mockImageModel,
+        prompt: 'A space station',
+        providerOptions: {
+          openai: {
+            quality: 'hd',
+            style: 'vivid'
+          }
+        }
+      })
+    })
+
+    it('should support custom headers', async () => {
+      await executor.generateImage({
+        model: 'dall-e-3',
+        prompt: 'A robot',
+        headers: {
+          'X-Custom-Header': 'test-value'
+        }
+      })
+
+      expect(aiGenerateImage).toHaveBeenCalledWith({
+        model: mockImageModel,
+        prompt: 'A robot',
+        headers: {
+          'X-Custom-Header': 'test-value'
+        }
+      })
+    })
+  })
+
+  describe('Plugin integration', () => {
+    it('should execute plugins in correct order', async () => {
+      const pluginCallOrder: string[] = []
+
+      const testPlugin: AiPlugin = {
+        name: 'test-plugin',
+        onRequestStart: vi.fn(async () => {
+          pluginCallOrder.push('onRequestStart')
+        }),
+        transformParams: vi.fn(async (params) => {
+          pluginCallOrder.push('transformParams')
+          return { ...params, size: '512x512' }
+        }),
+        transformResult: vi.fn(async (result) => {
+          pluginCallOrder.push('transformResult')
+          return { ...result, processed: true }
+        }),
+        onRequestEnd: vi.fn(async () => {
+          pluginCallOrder.push('onRequestEnd')
+        })
+      }
+
+      const executorWithPlugin = RuntimeExecutor.create(
+        'openai',
+        {
+          apiKey: 'test-key'
+        },
+        [testPlugin]
+      )
+
+      const result = await executorWithPlugin.generateImage({ model: 'dall-e-3', prompt: 'A test image' })
+
+      expect(pluginCallOrder).toEqual(['onRequestStart', 'transformParams', 'transformResult', 'onRequestEnd'])
+
+      expect(testPlugin.transformParams).toHaveBeenCalledWith(
+        { prompt: 'A test image' },
+        expect.objectContaining({
+          providerId: 'openai',
+          modelId: 'dall-e-3'
+        })
+      )
+
+      expect(aiGenerateImage).toHaveBeenCalledWith({
+        model: mockImageModel,
+        prompt: 'A test image',
+        size: '512x512' // Should be transformed by plugin
+      })
+
+      expect(result).toEqual({
+        ...mockGenerateImageResult,
+        processed: true // Should be transformed by plugin
+      })
+    })
+
+    it('should handle model resolution through plugins', async () => {
+      const customImageModel = {
+        modelId: 'custom-model',
+        provider: 'openai'
+      } as ImageModelV2
+
+      const modelResolutionPlugin: AiPlugin = {
+        name: 'model-resolver',
+        resolveModel: vi.fn(async () => customImageModel)
+      }
+
+      const executorWithPlugin = RuntimeExecutor.create(
+        'openai',
+        {
+          apiKey: 'test-key'
+        },
+        [modelResolutionPlugin]
+      )
+
+      await executorWithPlugin.generateImage({ model: 'dall-e-3', prompt: 'A test image' })
+
+      expect(modelResolutionPlugin.resolveModel).toHaveBeenCalledWith(
+        'dall-e-3',
+        expect.objectContaining({
+          providerId: 'openai',
+          modelId: 'dall-e-3'
+        })
+      )
+
+      expect(aiGenerateImage).toHaveBeenCalledWith({
+        model: customImageModel,
+        prompt: 'A test image'
+      })
+    })
+
+    it('should support recursive calls from plugins', async () => {
+      const recursivePlugin: AiPlugin = {
+        name: 'recursive-plugin',
+        transformParams: vi.fn(async (params, context) => {
+          if (!context.isRecursiveCall && params.prompt === 'original') {
+            // Make a recursive call with modified prompt
+            await context.recursiveCall({
+              model: 'dall-e-3',
+              prompt: 'modified'
+            })
+          }
+          return params
+        })
+      }
+
+      const executorWithPlugin = RuntimeExecutor.create(
+        'openai',
+        {
+          apiKey: 'test-key'
+        },
+        [recursivePlugin]
+      )
+
+      await executorWithPlugin.generateImage({ model: 'dall-e-3', prompt: 'original' })
+
+      expect(recursivePlugin.transformParams).toHaveBeenCalledTimes(2)
+      expect(aiGenerateImage).toHaveBeenCalledTimes(2)
+    })
+  })
+
+  describe('Error handling', () => {
+    it('should handle model creation errors', async () => {
+      const modelError = new Error('Failed to get image model')
+      vi.mocked(globalRegistryManagement.imageModel).mockImplementation(() => {
+        throw modelError
+      })
+
+      await expect(executor.generateImage({ model: 'invalid-model', prompt: 'A test image' })).rejects.toThrow(
+        ImageGenerationError
+      )
+    })
+
+    it('should handle ImageModelResolutionError correctly', async () => {
+      const resolutionError = new ImageModelResolutionError('invalid-model', 'openai', new Error('Model not found'))
+      vi.mocked(globalRegistryManagement.imageModel).mockImplementation(() => {
+        throw resolutionError
+      })
+
+      const thrownError = await executor
+        .generateImage({ model: 'invalid-model', prompt: 'A test image' })
+        .catch((error) => error)
+
+      expect(thrownError).toBeInstanceOf(ImageGenerationError)
+      expect(thrownError.message).toContain('Failed to generate image:')
+      expect(thrownError.providerId).toBe('openai')
+      expect(thrownError.modelId).toBe('invalid-model')
+      expect(thrownError.cause).toBeInstanceOf(ImageModelResolutionError)
+      expect(thrownError.cause.message).toContain('Failed to resolve image model: invalid-model')
+    })
+
+    it('should handle ImageModelResolutionError without provider', async () => {
+      const resolutionError = new ImageModelResolutionError('unknown-model')
+      vi.mocked(globalRegistryManagement.imageModel).mockImplementation(() => {
+        throw resolutionError
+      })
+
+      await expect(executor.generateImage({ model: 'unknown-model', prompt: 'A test image' })).rejects.toThrow(
+        ImageGenerationError
+      )
+    })
+
+    it('should handle image generation API errors', async () => {
+      const apiError = new Error('API request failed')
+      vi.mocked(aiGenerateImage).mockRejectedValue(apiError)
+
+      await expect(executor.generateImage({ model: 'dall-e-3', prompt: 'A test image' })).rejects.toThrow(
+        'Failed to generate image:'
+      )
+    })
+
+    it('should handle NoImageGeneratedError', async () => {
+      const noImageError = new NoImageGeneratedError({
+        cause: new Error('No image generated'),
+        responses: []
+      })
+
+      vi.mocked(aiGenerateImage).mockRejectedValue(noImageError)
+      vi.mocked(NoImageGeneratedError.isInstance).mockReturnValue(true)
+
+      await expect(executor.generateImage({ model: 'dall-e-3', prompt: 'A test image' })).rejects.toThrow(
+        'Failed to generate image:'
+      )
+    })
+
+    it('should execute onError plugin hook on failure', async () => {
+      const error = new Error('Generation failed')
+      vi.mocked(aiGenerateImage).mockRejectedValue(error)
+
+      const errorPlugin: AiPlugin = {
+        name: 'error-handler',
+        onError: vi.fn()
+      }
+
+      const executorWithPlugin = RuntimeExecutor.create(
+        'openai',
+        {
+          apiKey: 'test-key'
+        },
+        [errorPlugin]
+      )
+
+      await expect(executorWithPlugin.generateImage({ model: 'dall-e-3', prompt: 'A test image' })).rejects.toThrow(
+        'Failed to generate image:'
+      )
+
+      expect(errorPlugin.onError).toHaveBeenCalledWith(
+        error,
+        expect.objectContaining({
+          providerId: 'openai',
+          modelId: 'dall-e-3'
+        })
+      )
+    })
+
+    it('should handle abort signal timeout', async () => {
+      const abortError = new Error('Operation was aborted')
+      abortError.name = 'AbortError'
+      vi.mocked(aiGenerateImage).mockRejectedValue(abortError)
+
+      const abortController = new AbortController()
+      setTimeout(() => abortController.abort(), 10)
+
+      await expect(
+        executor.generateImage({ model: 'dall-e-3', prompt: 'A test image', abortSignal: abortController.signal })
+      ).rejects.toThrow('Failed to generate image:')
+    })
+  })
+
+  describe('Multiple providers support', () => {
+    it('should work with different providers', async () => {
+      const googleExecutor = RuntimeExecutor.create('google', {
+        apiKey: 'google-key'
+      })
+
+      await googleExecutor.generateImage({ model: 'imagen-3.0-generate-002', prompt: 'A landscape' })
+
+      expect(globalRegistryManagement.imageModel).toHaveBeenCalledWith('google|imagen-3.0-generate-002')
+    })
+
+    it('should support xAI Grok image models', async () => {
+      const xaiExecutor = RuntimeExecutor.create('xai', {
+        apiKey: 'xai-key'
+      })
+
+      await xaiExecutor.generateImage({ model: 'grok-2-image', prompt: 'A futuristic robot' })
+
+      expect(globalRegistryManagement.imageModel).toHaveBeenCalledWith('xai|grok-2-image')
+    })
+  })
+
+  describe('Advanced features', () => {
+    it('should support batch image generation with maxImagesPerCall', async () => {
+      await executor.generateImage({ model: 'dall-e-3', prompt: 'A test image', n: 10, maxImagesPerCall: 5 })
+
+      expect(aiGenerateImage).toHaveBeenCalledWith({
+        model: mockImageModel,
+        prompt: 'A test image',
+        n: 10,
+        maxImagesPerCall: 5
+      })
+    })
+
+    it('should support retries with maxRetries', async () => {
+      await executor.generateImage({ model: 'dall-e-3', prompt: 'A test image', maxRetries: 3 })
+
+      expect(aiGenerateImage).toHaveBeenCalledWith({
+        model: mockImageModel,
+        prompt: 'A test image',
+        maxRetries: 3
+      })
+    })
+
+    it('should handle warnings from the model', async () => {
+      const resultWithWarnings = {
+        ...mockGenerateImageResult,
+        warnings: [
+          {
+            type: 'unsupported-setting',
+            message: 'Size parameter not supported for this model'
+          }
+        ]
+      }
+
+      vi.mocked(aiGenerateImage).mockResolvedValue(resultWithWarnings)
+
+      const result = await executor.generateImage({
+        model: 'dall-e-3',
+        prompt: 'A test image',
+        size: '2048x2048' // Unsupported size
+      })
+
+      expect(result.warnings).toHaveLength(1)
+      expect(result.warnings[0].type).toBe('unsupported-setting')
+    })
+
+    it('should provide access to provider metadata', async () => {
+      const result = await executor.generateImage({ model: 'dall-e-3', prompt: 'A test image' })
+
+      expect(result.providerMetadata).toBeDefined()
+      expect(result.providerMetadata.openai).toBeDefined()
+    })
+
+    it('should provide response metadata', async () => {
+      const resultWithMetadata = {
+        ...mockGenerateImageResult,
+        responses: [
+          {
+            timestamp: new Date(),
+            modelId: 'dall-e-3',
+            headers: { 'x-request-id': 'test-123' }
+          }
+        ]
+      }
+
+      vi.mocked(aiGenerateImage).mockResolvedValue(resultWithMetadata)
+
+      const result = await executor.generateImage({ model: 'dall-e-3', prompt: 'A test image' })
+
+      expect(result.responses).toHaveLength(1)
+      expect(result.responses[0].modelId).toBe('dall-e-3')
+      expect(result.responses[0].headers).toEqual({ 'x-request-id': 'test-123' })
+    })
+  })
+})

packages/aiCore/src/core/runtime/errors.ts

@@ -0,0 +1,38 @@
+/**
+ * Error classes for runtime operations
+ */
+
+/**
+ * Error thrown when image generation fails
+ */
+export class ImageGenerationError extends Error {
+  constructor(
+    message: string,
+    public providerId?: string,
+    public modelId?: string,
+    public cause?: Error
+  ) {
+    super(message)
+    this.name = 'ImageGenerationError'
+
+    // Maintain proper stack trace (for V8 engines)
+    if (Error.captureStackTrace) {
+      Error.captureStackTrace(this, ImageGenerationError)
+    }
+  }
+}
+
+/**
+ * Error thrown when model resolution fails during image generation
+ */
+export class ImageModelResolutionError extends ImageGenerationError {
+  constructor(modelId: string, providerId?: string, cause?: Error) {
+    super(
+      `Failed to resolve image model: ${modelId}${providerId ? ` for provider: ${providerId}` : ''}`,
+      providerId,
+      modelId,
+      cause
+    )
+    this.name = 'ImageModelResolutionError'
+  }
+}

packages/aiCore/src/core/runtime/executor.ts

@@ -0,0 +1,310 @@
+/**
+ * 运行时执行器
+ * 专注于插件化的AI调用处理
+ */
+import { ImageModelV2, LanguageModelV2, LanguageModelV2Middleware } from '@ai-sdk/provider'
+import {
+  experimental_generateImage as _generateImage,
+  generateObject as _generateObject,
+  generateText as _generateText,
+  LanguageModel,
+  streamObject as _streamObject,
+  streamText as _streamText
+} from 'ai'
+
+import { globalModelResolver } from '../models'
+import { type ModelConfig } from '../models/types'
+import { type AiPlugin, type AiRequestContext, definePlugin } from '../plugins'
+import { type ProviderId } from '../providers'
+import { ImageGenerationError, ImageModelResolutionError } from './errors'
+import { PluginEngine } from './pluginEngine'
+import type {
+  generateImageParams,
+  generateObjectParams,
+  generateTextParams,
+  RuntimeConfig,
+  streamObjectParams,
+  streamTextParams
+} from './types'
+
+export class RuntimeExecutor<T extends ProviderId = ProviderId> {
+  public pluginEngine: PluginEngine<T>
+  // private options: ProviderSettingsMap[T]
+  private config: RuntimeConfig<T>
+
+  constructor(config: RuntimeConfig<T>) {
+    // if (!isProviderSupported(config.providerId)) {
+    //   throw new Error(`Unsupported provider: ${config.providerId}`)
+    // }
+
+    // 存储options供后续使用
+    // this.options = config.options
+    this.config = config
+    // 创建插件客户端
+    this.pluginEngine = new PluginEngine(config.providerId, config.plugins || [])
+  }
+
+  private createResolveModelPlugin(middlewares?: LanguageModelV2Middleware[]) {
+    return definePlugin({
+      name: '_internal_resolveModel',
+      enforce: 'post',
+
+      resolveModel: async (modelId: string) => {
+        // 注意：extraModelConfig 暂时不支持，已在新架构中移除
+        return await this.resolveModel(modelId, middlewares)
+      }
+    })
+  }
+
+  private createResolveImageModelPlugin() {
+    return definePlugin({
+      name: '_internal_resolveImageModel',
+      enforce: 'post',
+
+      resolveModel: async (modelId: string) => {
+        return await this.resolveImageModel(modelId)
+      }
+    })
+  }
+
+  private createConfigureContextPlugin() {
+    return definePlugin({
+      name: '_internal_configureContext',
+      configureContext: async (context: AiRequestContext) => {
+        context.executor = this
+      }
+    })
+  }
+
+  // === 高阶重载：直接使用模型 ===
+
+  /**
+   * 流式文本生成
+   */
+  async streamText(
+    params: streamTextParams,
+    options?: {
+      middlewares?: LanguageModelV2Middleware[]
+    }
+  ): Promise<ReturnType<typeof _streamText>> {
+    const { model } = params
+
+    // 根据 model 类型决定插件配置
+    if (typeof model === 'string') {
+      this.pluginEngine.usePlugins([
+        this.createResolveModelPlugin(options?.middlewares),
+        this.createConfigureContextPlugin()
+      ])
+    } else {
+      this.pluginEngine.usePlugins([this.createConfigureContextPlugin()])
+    }
+
+    return this.pluginEngine.executeStreamWithPlugins(
+      'streamText',
+      params,
+      (resolvedModel, transformedParams, streamTransforms) => {
+        const experimental_transform =
+          params?.experimental_transform ?? (streamTransforms.length > 0 ? streamTransforms : undefined)
+
+        return _streamText({
+          ...transformedParams,
+          model: resolvedModel,
+          experimental_transform
+        })
+      }
+    )
+  }
+
+  // === 其他方法的重载 ===
+
+  /**
+   * 生成文本
+   */
+  async generateText(
+    params: generateTextParams,
+    options?: {
+      middlewares?: LanguageModelV2Middleware[]
+    }
+  ): Promise<ReturnType<typeof _generateText>> {
+    const { model } = params
+
+    // 根据 model 类型决定插件配置
+    if (typeof model === 'string') {
+      this.pluginEngine.usePlugins([
+        this.createResolveModelPlugin(options?.middlewares),
+        this.createConfigureContextPlugin()
+      ])
+    } else {
+      this.pluginEngine.usePlugins([this.createConfigureContextPlugin()])
+    }
+
+    return this.pluginEngine.executeWithPlugins<Parameters<typeof _generateText>[0], ReturnType<typeof _generateText>>(
+      'generateText',
+      params,
+      (resolvedModel, transformedParams) => _generateText({ ...transformedParams, model: resolvedModel })
+    )
+  }
+
+  /**
+   * 生成结构化对象
+   */
+  async generateObject(
+    params: generateObjectParams,
+    options?: {
+      middlewares?: LanguageModelV2Middleware[]
+    }
+  ): Promise<ReturnType<typeof _generateObject>> {
+    const { model } = params
+
+    // 根据 model 类型决定插件配置
+    if (typeof model === 'string') {
+      this.pluginEngine.usePlugins([
+        this.createResolveModelPlugin(options?.middlewares),
+        this.createConfigureContextPlugin()
+      ])
+    } else {
+      this.pluginEngine.usePlugins([this.createConfigureContextPlugin()])
+    }
+
+    return this.pluginEngine.executeWithPlugins<generateObjectParams, ReturnType<typeof _generateObject>>(
+      'generateObject',
+      params,
+      async (resolvedModel, transformedParams) => _generateObject({ ...transformedParams, model: resolvedModel })
+    )
+  }
+
+  /**
+   * 流式生成结构化对象
+   */
+  streamObject(
+    params: streamObjectParams,
+    options?: {
+      middlewares?: LanguageModelV2Middleware[]
+    }
+  ): Promise<ReturnType<typeof _streamObject>> {
+    const { model } = params
+
+    // 根据 model 类型决定插件配置
+    if (typeof model === 'string') {
+      this.pluginEngine.usePlugins([
+        this.createResolveModelPlugin(options?.middlewares),
+        this.createConfigureContextPlugin()
+      ])
+    } else {
+      this.pluginEngine.usePlugins([this.createConfigureContextPlugin()])
+    }
+
+    return this.pluginEngine.executeStreamWithPlugins('streamObject', params, (resolvedModel, transformedParams) =>
+      _streamObject({ ...transformedParams, model: resolvedModel })
+    )
+  }
+
+  /**
+   * 生成图像
+   */
+  generateImage(params: generateImageParams): Promise<ReturnType<typeof _generateImage>> {
+    try {
+      const { model } = params
+
+      // 根据 model 类型决定插件配置
+      if (typeof model === 'string') {
+        this.pluginEngine.usePlugins([this.createResolveImageModelPlugin(), this.createConfigureContextPlugin()])
+      } else {
+        this.pluginEngine.usePlugins([this.createConfigureContextPlugin()])
+      }
+
+      return this.pluginEngine.executeImageWithPlugins('generateImage', params, (resolvedModel, transformedParams) =>
+        _generateImage({ ...transformedParams, model: resolvedModel })
+      )
+    } catch (error) {
+      if (error instanceof Error) {
+        const modelId = typeof params.model === 'string' ? params.model : params.model.modelId
+        throw new ImageGenerationError(
+          `Failed to generate image: ${error.message}`,
+          this.config.providerId,
+          modelId,
+          error
+        )
+      }
+      throw error
+    }
+  }
+
+  // === 辅助方法 ===
+
+  /**
+   * 解析模型：如果是字符串则创建模型，如果是模型则直接返回
+   */
+  private async resolveModel(
+    modelOrId: LanguageModel,
+    middlewares?: LanguageModelV2Middleware[]
+  ): Promise<LanguageModelV2> {
+    if (typeof modelOrId === 'string') {
+      // 🎯 字符串modelId，使用新的ModelResolver解析，传递完整参数
+      return await globalModelResolver.resolveLanguageModel(
+        modelOrId, // 支持 'gpt-4' 和 'aihubmix:anthropic:claude-3.5-sonnet'
+        this.config.providerId, // fallback provider
+        this.config.providerSettings, // provider options
+        middlewares // 中间件数组
+      )
+    } else {
+      // 已经是模型，直接返回
+      return modelOrId
+    }
+  }
+
+  /**
+   * 解析图像模型：如果是字符串则创建图像模型，如果是模型则直接返回
+   */
+  private async resolveImageModel(modelOrId: ImageModelV2 | string): Promise<ImageModelV2> {
+    try {
+      if (typeof modelOrId === 'string') {
+        // 字符串modelId，使用新的ModelResolver解析
+        return await globalModelResolver.resolveImageModel(
+          modelOrId, // 支持 'dall-e-3' 和 'aihubmix:openai:dall-e-3'
+          this.config.providerId // fallback provider
+        )
+      } else {
+        // 已经是模型，直接返回
+        return modelOrId
+      }
+    } catch (error) {
+      throw new ImageModelResolutionError(
+        typeof modelOrId === 'string' ? modelOrId : modelOrId.modelId,
+        this.config.providerId,
+        error instanceof Error ? error : undefined
+      )
+    }
+  }
+
+  // === 静态工厂方法 ===
+
+  /**
+   * 创建执行器 - 支持已知provider的类型安全
+   */
+  static create<T extends ProviderId>(
+    providerId: T,
+    options: ModelConfig<T>['providerSettings'],
+    plugins?: AiPlugin[]
+  ): RuntimeExecutor<T> {
+    return new RuntimeExecutor({
+      providerId,
+      providerSettings: options,
+      plugins
+    })
+  }
+
+  /**
+   * 创建OpenAI Compatible执行器
+   */
+  static createOpenAICompatible(
+    options: ModelConfig<'openai-compatible'>['providerSettings'],
+    plugins: AiPlugin[] = []
+  ): RuntimeExecutor<'openai-compatible'> {
+    return new RuntimeExecutor({
+      providerId: 'openai-compatible',
+      providerSettings: options,
+      plugins
+    })
+  }
+}

packages/aiCore/src/core/runtime/index.ts

@@ -0,0 +1,117 @@
+/**
+ * Runtime 模块导出
+ * 专注于运行时插件化AI调用处理
+ */
+
+// 主要的运行时执行器
+export { RuntimeExecutor } from './executor'
+
+// 导出类型
+export type { RuntimeConfig } from './types'
+
+// === 便捷工厂函数 ===
+
+import { LanguageModelV2Middleware } from '@ai-sdk/provider'
+
+import { type AiPlugin } from '../plugins'
+import { type ProviderId, type ProviderSettingsMap } from '../providers/types'
+import { RuntimeExecutor } from './executor'
+
+/**
+ * 创建运行时执行器 - 支持类型安全的已知provider
+ */
+export function createExecutor<T extends ProviderId>(
+  providerId: T,
+  options: ProviderSettingsMap[T] & { mode?: 'chat' | 'responses' },
+  plugins?: AiPlugin[]
+): RuntimeExecutor<T> {
+  return RuntimeExecutor.create(providerId, options, plugins)
+}
+
+/**
+ * 创建OpenAI Compatible执行器
+ */
+export function createOpenAICompatibleExecutor(
+  options: ProviderSettingsMap['openai-compatible'] & { mode?: 'chat' | 'responses' },
+  plugins: AiPlugin[] = []
+): RuntimeExecutor<'openai-compatible'> {
+  return RuntimeExecutor.createOpenAICompatible(options, plugins)
+}
+
+// === 直接调用API（无需创建executor实例）===
+
+/**
+ * 直接流式文本生成 - 支持middlewares
+ */
+export async function streamText<T extends ProviderId>(
+  providerId: T,
+  options: ProviderSettingsMap[T] & { mode?: 'chat' | 'responses' },
+  params: Parameters<RuntimeExecutor<T>['streamText']>[0],
+  plugins?: AiPlugin[],
+  middlewares?: LanguageModelV2Middleware[]
+): Promise<ReturnType<RuntimeExecutor<T>['streamText']>> {
+  const executor = createExecutor(providerId, options, plugins)
+  return executor.streamText(params, { middlewares })
+}
+
+/**
+ * 直接生成文本 - 支持middlewares
+ */
+export async function generateText<T extends ProviderId>(
+  providerId: T,
+  options: ProviderSettingsMap[T] & { mode?: 'chat' | 'responses' },
+  params: Parameters<RuntimeExecutor<T>['generateText']>[0],
+  plugins?: AiPlugin[],
+  middlewares?: LanguageModelV2Middleware[]
+): Promise<ReturnType<RuntimeExecutor<T>['generateText']>> {
+  const executor = createExecutor(providerId, options, plugins)
+  return executor.generateText(params, { middlewares })
+}
+
+/**
+ * 直接生成结构化对象 - 支持middlewares
+ */
+export async function generateObject<T extends ProviderId>(
+  providerId: T,
+  options: ProviderSettingsMap[T] & { mode?: 'chat' | 'responses' },
+  params: Parameters<RuntimeExecutor<T>['generateObject']>[0],
+  plugins?: AiPlugin[],
+  middlewares?: LanguageModelV2Middleware[]
+): Promise<ReturnType<RuntimeExecutor<T>['generateObject']>> {
+  const executor = createExecutor(providerId, options, plugins)
+  return executor.generateObject(params, { middlewares })
+}
+
+/**
+ * 直接流式生成结构化对象 - 支持middlewares
+ */
+export async function streamObject<T extends ProviderId>(
+  providerId: T,
+  options: ProviderSettingsMap[T] & { mode?: 'chat' | 'responses' },
+  params: Parameters<RuntimeExecutor<T>['streamObject']>[0],
+  plugins?: AiPlugin[],
+  middlewares?: LanguageModelV2Middleware[]
+): Promise<ReturnType<RuntimeExecutor<T>['streamObject']>> {
+  const executor = createExecutor(providerId, options, plugins)
+  return executor.streamObject(params, { middlewares })
+}
+
+/**
+ * 直接生成图像 - 支持middlewares
+ */
+export async function generateImage<T extends ProviderId>(
+  providerId: T,
+  options: ProviderSettingsMap[T] & { mode?: 'chat' | 'responses' },
+  params: Parameters<RuntimeExecutor<T>['generateImage']>[0],
+  plugins?: AiPlugin[]
+): Promise<ReturnType<RuntimeExecutor<T>['generateImage']>> {
+  const executor = createExecutor(providerId, options, plugins)
+  return executor.generateImage(params)
+}
+
+// === Agent 功能预留 ===
+// 未来将在 ../agents/ 文件夹中添加：
+// - AgentExecutor.ts
+// - WorkflowManager.ts
+// - ConversationManager.ts
+// 并在此处导出相关API

packages/aiCore/src/core/runtime/pluginEngine.ts

@@ -0,0 +1,296 @@
+/* eslint-disable @eslint-react/naming-convention/context-name */
+import { ImageModelV2 } from '@ai-sdk/provider'
+import { experimental_generateImage, generateObject, generateText, LanguageModel, streamObject, streamText } from 'ai'
+
+import { type AiPlugin, createContext, PluginManager } from '../plugins'
+import { type ProviderId } from '../providers/types'
+
+/**
+ * 插件增强的 AI 客户端
+ * 专注于插件处理，不暴露用户API
+ */
+export class PluginEngine<T extends ProviderId = ProviderId> {
+  private pluginManager: PluginManager
+
+  constructor(
+    private readonly providerId: T,
+    // private readonly options: ProviderSettingsMap[T],
+    plugins: AiPlugin[] = []
+  ) {
+    this.pluginManager = new PluginManager(plugins)
+  }
+
+  /**
+   * 添加插件
+   */
+  use(plugin: AiPlugin): this {
+    this.pluginManager.use(plugin)
+    return this
+  }
+
+  /**
+   * 批量添加插件
+   */
+  usePlugins(plugins: AiPlugin[]): this {
+    plugins.forEach((plugin) => this.use(plugin))
+    return this
+  }
+
+  /**
+   * 移除插件
+   */
+  removePlugin(pluginName: string): this {
+    this.pluginManager.remove(pluginName)
+    return this
+  }
+
+  /**
+   * 获取插件统计
+   */
+  getPluginStats() {
+    return this.pluginManager.getStats()
+  }
+
+  /**
+   * 获取所有插件
+   */
+  getPlugins() {
+    return this.pluginManager.getPlugins()
+  }
+
+  /**
+   * 执行带插件的操作（非流式）
+   * 提供给AiExecutor使用
+   */
+  async executeWithPlugins<
+    TParams extends Parameters<typeof generateText | typeof generateObject>[0],
+    TResult extends ReturnType<typeof generateText | typeof generateObject>
+  >(
+    methodName: string,
+    params: TParams,
+    executor: (model: LanguageModel, transformedParams: TParams) => TResult,
+    _context?: ReturnType<typeof createContext>
+  ): Promise<TResult> {
+    // 统一处理模型解析
+    let resolvedModel: LanguageModel | undefined
+    let modelId: string
+    const { model } = params
+    if (typeof model === 'string') {
+      // 字符串：需要通过插件解析
+      modelId = model
+    } else {
+      // 模型对象：直接使用
+      resolvedModel = model
+      modelId = model.modelId
+    }
+
+    // 使用正确的createContext创建请求上下文
+    const context = _context ? _context : createContext(this.providerId, model, params)
+
+    // 🔥 为上下文添加递归调用能力
+    context.recursiveCall = async (newParams: any): Promise<TResult> => {
+      // 递归调用自身，重新走完整的插件流程
+      context.isRecursiveCall = true
+      const result = await this.executeWithPlugins(methodName, newParams, executor, context)
+      context.isRecursiveCall = false
+      return result
+    }
+
+    try {
+      // 0. 配置上下文
+      await this.pluginManager.executeConfigureContext(context)
+
+      // 1. 触发请求开始事件
+      await this.pluginManager.executeParallel('onRequestStart', context)
+
+      // 2. 解析模型（如果是字符串）
+      if (typeof model === 'string') {
+        const resolved = await this.pluginManager.executeFirst<LanguageModel>('resolveModel', modelId, context)
+        if (!resolved) {
+          throw new Error(`Failed to resolve model: ${modelId}`)
+        }
+        resolvedModel = resolved
+      }
+
+      if (!resolvedModel) {
+        throw new Error(`Model resolution failed: no model available`)
+      }
+
+      // 3. 转换请求参数
+      const transformedParams = await this.pluginManager.executeSequential('transformParams', params, context)
+
+      // 4. 执行具体的 API 调用
+      const result = await executor(resolvedModel, transformedParams)
+
+      // 5. 转换结果（对于非流式调用）
+      const transformedResult = await this.pluginManager.executeSequential('transformResult', result, context)
+
+      // 6. 触发完成事件
+      await this.pluginManager.executeParallel('onRequestEnd', context, transformedResult)
+
+      return transformedResult
+    } catch (error) {
+      // 7. 触发错误事件
+      await this.pluginManager.executeParallel('onError', context, undefined, error as Error)
+      throw error
+    }
+  }
+
+  /**
+   * 执行带插件的图像生成操作
+   * 提供给AiExecutor使用
+   */
+  async executeImageWithPlugins<
+    TParams extends Omit<Parameters<typeof experimental_generateImage>[0], 'model'> & { model: string | ImageModelV2 },
+    TResult extends ReturnType<typeof experimental_generateImage>
+  >(
+    methodName: string,
+    params: TParams,
+    executor: (model: ImageModelV2, transformedParams: TParams) => TResult,
+    _context?: ReturnType<typeof createContext>
+  ): Promise<TResult> {
+    // 统一处理模型解析
+    let resolvedModel: ImageModelV2 | undefined
+    let modelId: string
+    const { model } = params
+    if (typeof model === 'string') {
+      // 字符串：需要通过插件解析
+      modelId = model
+    } else {
+      // 模型对象：直接使用
+      resolvedModel = model
+      modelId = model.modelId
+    }
+
+    // 使用正确的createContext创建请求上下文
+    const context = _context ? _context : createContext(this.providerId, model, params)
+
+    // 🔥 为上下文添加递归调用能力
+    context.recursiveCall = async (newParams: any): Promise<TResult> => {
+      // 递归调用自身，重新走完整的插件流程
+      context.isRecursiveCall = true
+      const result = await this.executeImageWithPlugins(methodName, newParams, executor, context)
+      context.isRecursiveCall = false
+      return result
+    }
+
+    try {
+      // 0. 配置上下文
+      await this.pluginManager.executeConfigureContext(context)
+
+      // 1. 触发请求开始事件
+      await this.pluginManager.executeParallel('onRequestStart', context)
+
+      // 2. 解析模型（如果是字符串）
+      if (typeof model === 'string') {
+        const resolved = await this.pluginManager.executeFirst<ImageModelV2>('resolveModel', modelId, context)
+        if (!resolved) {
+          throw new Error(`Failed to resolve image model: ${modelId}`)
+        }
+        resolvedModel = resolved
+      }
+
+      if (!resolvedModel) {
+        throw new Error(`Image model resolution failed: no model available`)
+      }
+
+      // 3. 转换请求参数
+      const transformedParams = await this.pluginManager.executeSequential('transformParams', params, context)
+
+      // 4. 执行具体的 API 调用
+      const result = await executor(resolvedModel, transformedParams)
+
+      // 5. 转换结果
+      const transformedResult = await this.pluginManager.executeSequential('transformResult', result, context)
+
+      // 6. 触发完成事件
+      await this.pluginManager.executeParallel('onRequestEnd', context, transformedResult)
+
+      return transformedResult
+    } catch (error) {
+      // 7. 触发错误事件
+      await this.pluginManager.executeParallel('onError', context, undefined, error as Error)
+      throw error
+    }
+  }
+
+  /**
+   * 执行流式调用的通用逻辑（支持流转换器）
+   * 提供给AiExecutor使用
+   */
+  async executeStreamWithPlugins<
+    TParams extends Parameters<typeof streamText | typeof streamObject>[0],
+    TResult extends ReturnType<typeof streamText | typeof streamObject>
+  >(
+    methodName: string,
+    params: TParams,
+    executor: (model: LanguageModel, transformedParams: TParams, streamTransforms: any[]) => TResult,
+    _context?: ReturnType<typeof createContext>
+  ): Promise<TResult> {
+    // 统一处理模型解析
+    let resolvedModel: LanguageModel | undefined
+    let modelId: string
+    const { model } = params
+    if (typeof model === 'string') {
+      // 字符串：需要通过插件解析
+      modelId = model
+    } else {
+      // 模型对象：直接使用
+      resolvedModel = model
+      modelId = model.modelId
+    }
+
+    // 创建请求上下文
+    const context = _context ? _context : createContext(this.providerId, model, params)
+
+    // 🔥 为上下文添加递归调用能力
+    context.recursiveCall = async (newParams: any): Promise<TResult> => {
+      // 递归调用自身，重新走完整的插件流程
+      context.isRecursiveCall = true
+      const result = await this.executeStreamWithPlugins(methodName, newParams, executor, context)
+      context.isRecursiveCall = false
+      return result
+    }
+
+    try {
+      // 0. 配置上下文
+      await this.pluginManager.executeConfigureContext(context)
+
+      // 1. 触发请求开始事件
+      await this.pluginManager.executeParallel('onRequestStart', context)
+
+      // 2. 解析模型（如果是字符串）
+      if (typeof model === 'string') {
+        const resolved = await this.pluginManager.executeFirst<LanguageModel>('resolveModel', modelId, context)
+        if (!resolved) {
+          throw new Error(`Failed to resolve model: ${modelId}`)
+        }
+        resolvedModel = resolved
+      }
+
+      if (!resolvedModel) {
+        throw new Error(`Model resolution failed: no model available`)
+      }
+
+      // 3. 转换请求参数
+      const transformedParams = await this.pluginManager.executeSequential('transformParams', params, context)
+
+      // 4. 收集流转换器
+      const streamTransforms = this.pluginManager.collectStreamTransforms(transformedParams, context)
+
+      // 5. 执行流式 API 调用
+      const result = await executor(resolvedModel, transformedParams, streamTransforms)
+
+      const transformedResult = await this.pluginManager.executeSequential('transformResult', result, context)
+
+      // 6. 触发完成事件（注意：对于流式调用，这里触发的是开始流式响应的事件）
+      await this.pluginManager.executeParallel('onRequestEnd', context, transformedResult)
+
+      return transformedResult
+    } catch (error) {
+      // 7. 触发错误事件
+      await this.pluginManager.executeParallel('onError', context, undefined, error as Error)
+      throw error
+    }
+  }
+}

packages/aiCore/src/core/runtime/types.ts

@@ -0,0 +1,26 @@
+/**
+ * Runtime 层类型定义
+ */
+import { ImageModelV2 } from '@ai-sdk/provider'
+import { experimental_generateImage, generateObject, generateText, streamObject, streamText } from 'ai'
+
+import { type ModelConfig } from '../models/types'
+import { type AiPlugin } from '../plugins'
+import { type ProviderId } from '../providers/types'
+
+/**
+ * 运行时执行器配置
+ */
+export interface RuntimeConfig<T extends ProviderId = ProviderId> {
+  providerId: T
+  providerSettings: ModelConfig<T>['providerSettings'] & { mode?: 'chat' | 'responses' }
+  plugins?: AiPlugin[]
+}
+
+export type generateImageParams = Omit<Parameters<typeof experimental_generateImage>[0], 'model'> & {
+  model: string | ImageModelV2
+}
+export type generateObjectParams = Parameters<typeof generateObject>[0]
+export type generateTextParams = Parameters<typeof generateText>[0]
+export type streamObjectParams = Parameters<typeof streamObject>[0]
+export type streamTextParams = Parameters<typeof streamText>[0]

packages/aiCore/src/index.ts

@@ -0,0 +1,46 @@
+/**
+ * Cherry Studio AI Core Package
+ * 基于 Vercel AI SDK 的统一 AI Provider 接口
+ */
+
+// 导入内部使用的类和函数
+
+// ==================== 主要用户接口 ====================
+export {
+  createExecutor,
+  createOpenAICompatibleExecutor,
+  generateImage,
+  generateObject,
+  generateText,
+  streamText
+} from './core/runtime'
+
+// ==================== 高级API ====================
+export { globalModelResolver as modelResolver } from './core/models'
+
+// ==================== 插件系统 ====================
+export type { AiPlugin, AiRequestContext, HookResult, PluginManagerConfig } from './core/plugins'
+export { createContext, definePlugin, PluginManager } from './core/plugins'
+// export { createPromptToolUsePlugin, webSearchPlugin } from './core/plugins/built-in'
+export { PluginEngine } from './core/runtime/pluginEngine'
+
+// ==================== AI SDK 常用类型导出 ====================
+// 直接导出 AI SDK 的常用类型，方便使用
+export type { LanguageModelV2Middleware, LanguageModelV2StreamPart } from '@ai-sdk/provider'
+export type { ToolCall } from '@ai-sdk/provider-utils'
+export type { ReasoningPart } from '@ai-sdk/provider-utils'
+
+// ==================== 选项 ====================
+export {
+  createAnthropicOptions,
+  createGoogleOptions,
+  createOpenAIOptions,
+  type ExtractProviderOptions,
+  mergeProviderOptions,
+  type ProviderOptionsMap,
+  type TypedProviderOptions
+} from './core/options'
+
+// ==================== 包信息 ====================
+export const AI_CORE_VERSION = '1.0.0'
+export const AI_CORE_NAME = '@cherrystudio/ai-core'

packages/aiCore/src/types.ts

@@ -0,0 +1,2 @@
+// 重新导出插件类型
+export type { AiPlugin, AiRequestContext, HookResult, PluginManagerConfig } from './core/plugins/types'

packages/aiCore/tsconfig.json

@@ -0,0 +1,21 @@
+{
+  "compilerOptions": {
+    "allowSyntheticDefaultImports": true,
+    "declaration": true,
+    "emitDecoratorMetadata": true,
+    "esModuleInterop": true,
+    "experimentalDecorators": true,
+    "forceConsistentCasingInFileNames": true,
+    "module": "ESNext",
+    "moduleResolution": "bundler",
+    "noEmitOnError": false,
+    "outDir": "./dist",
+    "resolveJsonModule": true,
+    "rootDir": "./src",
+    "skipLibCheck": true,
+    "strict": true,
+    "target": "ES2020"
+  },
+  "exclude": ["node_modules", "dist"],
+  "include": ["src/**/*"]
+}

packages/aiCore/tsdown.config.ts

@@ -0,0 +1,14 @@
+import { defineConfig } from 'tsdown'
+
+export default defineConfig({
+  entry: {
+    index: 'src/index.ts',
+    'built-in/plugins/index': 'src/core/plugins/built-in/index.ts',
+    'provider/index': 'src/core/providers/index.ts'
+  },
+  outDir: 'dist',
+  format: ['esm', 'cjs'],
+  clean: true,
+  dts: true,
+  tsconfig: 'tsconfig.json'
+})

packages/aiCore/vitest.config.ts

@@ -0,0 +1,15 @@
+import { defineConfig } from 'vitest/config'
+
+export default defineConfig({
+  test: {
+    globals: true
+  },
+  resolve: {
+    alias: {
+      '@': './src'
+    }
+  },
+  esbuild: {
+    target: 'node18'
+  }
+})

packages/extension-table-plus/package.json

@@ -67,13 +67,13 @@
     "dist"
   ],
   "devDependencies": {
+    "@biomejs/biome": "2.2.4",
     "@tiptap/core": "^3.2.0",
     "@tiptap/pm": "^3.2.0",
     "eslint": "^9.22.0",
     "eslint-plugin-react-hooks": "^5.2.0",
     "eslint-plugin-simple-import-sort": "^12.1.1",
     "eslint-plugin-unused-imports": "^4.1.4",
-    "prettier": "^3.5.3",
     "tsdown": "^0.13.3"
   },
   "peerDependencies": {
@@ -87,7 +87,7 @@
   },
   "scripts": {
     "build": "tsdown",
-    "lint": "prettier ./src/ --write && eslint --fix ./src/"
+    "lint": "biome format ./src/ --write && eslint --fix ./src/"
   },
   "packageManager": "yarn@4.9.1"
 }

packages/shared/IpcChannel.ts

@@ -35,8 +35,10 @@ export enum IpcChannel {
   App_InstallBunBinary = 'app:install-bun-binary',
   App_LogToMain = 'app:log-to-main',
   App_SaveData = 'app:save-data',
+  App_GetDiskInfo = 'app:get-disk-info',
   App_SetFullScreen = 'app:set-full-screen',
   App_IsFullScreen = 'app:is-full-screen',
+  App_GetSystemFonts = 'app:get-system-fonts',
 
   App_MacIsProcessTrusted = 'app:mac-is-process-trusted',
   App_MacRequestProcessTrust = 'app:mac-request-process-trust',
@@ -83,10 +85,13 @@ export enum IpcChannel {
   Mcp_UploadDxt = 'mcp:upload-dxt',
   Mcp_AbortTool = 'mcp:abort-tool',
   Mcp_GetServerVersion = 'mcp:get-server-version',
-
+  Mcp_Progress = 'mcp:progress',
   // Python
   Python_Execute = 'python:execute',
 
+  // agent messages
+  AgentMessage_PersistExchange = 'agent-message:persist-exchange',
+
   //copilot
   Copilot_GetAuthMessage = 'copilot:get-auth-message',
   Copilot_GetCopilotToken = 'copilot:get-copilot-token',
@@ -123,6 +128,12 @@ export enum IpcChannel {
   Windows_SetMinimumSize = 'window:set-minimum-size',
   Windows_Resize = 'window:resize',
   Windows_GetSize = 'window:get-size',
+  Windows_Minimize = 'window:minimize',
+  Windows_Maximize = 'window:maximize',
+  Windows_Unmaximize = 'window:unmaximize',
+  Windows_Close = 'window:close',
+  Windows_IsMaximized = 'window:is-maximized',
+  Windows_MaximizedChanged = 'window:maximized-changed',
 
   KnowledgeBase_Create = 'knowledge-base:create',
   KnowledgeBase_Reset = 'knowledge-base:reset',
@@ -296,6 +307,21 @@ export enum IpcChannel {
   TRACE_CLEAN_LOCAL_DATA = 'trace:cleanLocalData',
   TRACE_ADD_STREAM_MESSAGE = 'trace:addStreamMessage',
 
+  // API Server
+  ApiServer_Start = 'api-server:start',
+  ApiServer_Stop = 'api-server:stop',
+  ApiServer_Restart = 'api-server:restart',
+  ApiServer_GetStatus = 'api-server:get-status',
+  ApiServer_GetConfig = 'api-server:get-config',
+
+  // Anthropic OAuth
+  Anthropic_StartOAuthFlow = 'anthropic:start-oauth-flow',
+  Anthropic_CompleteOAuthWithCode = 'anthropic:complete-oauth-with-code',
+  Anthropic_CancelOAuthFlow = 'anthropic:cancel-oauth-flow',
+  Anthropic_GetAccessToken = 'anthropic:get-access-token',
+  Anthropic_HasCredentials = 'anthropic:has-credentials',
+  Anthropic_ClearCredentials = 'anthropic:clear-credentials',
+
   // CodeTools
   CodeTools_Run = 'code-tools:run',
 

packages/shared/config/constant.ts

@@ -216,5 +216,6 @@ export enum codeTools {
   qwenCode = 'qwen-code',
   claudeCode = 'claude-code',
   geminiCli = 'gemini-cli',
-  openaiCodex = 'openai-codex'
+  openaiCodex = 'openai-codex',
+  iFlowCli = 'iflow-cli'
 }

packages/shared/config/types.ts

@@ -7,7 +7,7 @@ export type LoaderReturn = {
   loaderType: string
   status?: ProcessingStatus
   message?: string
-  messageSource?: 'preprocess' | 'embedding'
+  messageSource?: 'preprocess' | 'embedding' | 'validation'
 }
 
 export type FileChangeEventType = 'add' | 'change' | 'unlink' | 'addDir' | 'unlinkDir'
@@ -17,3 +17,8 @@ export type FileChangeEvent = {
   filePath: string
   watchPath: string
 }
+
+export type MCPProgressEvent = {
+  callId: string
+  progress: number // 0-1 range
+}

packages/shared/utils.ts

@@ -0,0 +1,6 @@
+export const defaultAppHeaders = () => {
+  return {
+    'HTTP-Referer': 'https://cherry-ai.com',
+    'X-Title': 'Cherry Studio'
+  }
+}

resources/cherry-studio/license.html

@@ -1,199 +1,274 @@
-<!DOCTYPE html>
+<!doctype html>
 <html lang="zh-CN">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>许可协议 | License Agreement</title>
+    <script src="https://cdn.tailwindcss.com"></script>
+  </head>
 
-<head>
+  <body class="bg-gray-50">
-  <meta charset="UTF-8">
+    <div class="mx-auto max-w-4xl px-4 py-8">
-  <meta name="viewport" content="width=device-width, initial-scale=1.0">
+      <!-- 中文版本 -->
-  <title>许可协议 | License Agreement</title>
+      <div class="mb-12">
-  <script src="https://cdn.tailwindcss.com"></script>
+        <h1 class="mb-8 text-3xl font-bold text-gray-900">许可协议</h1>
-</head>
 
-<body class="bg-gray-50">
+        <p class="mb-6 text-gray-700">
-  <div class="max-w-4xl mx-auto px-4 py-8">
+          本项目采用<strong>区分用户的双重许可 (User-Segmented Dual Licensing)</strong> 模式。
-    <!-- 中文版本 -->
-    <div class="mb-12">
-      <h1 class="text-3xl font-bold mb-8 text-gray-900">许可协议</h1>
-
-      <p class="mb-6 text-gray-700">本项目采用<strong>区分用户的双重许可 (User-Segmented Dual Licensing)</strong> 模式。</p>
-
-      <section class="mb-8">
-        <h2 class="text-xl font-semibold mb-4 text-gray-900">核心原则</h2>
-        <ul class="list-disc pl-6 space-y-2 text-gray-700">
-          <li><strong>个人用户 和 10人及以下企业/组织:</strong> 默认适用 <strong>GNU Affero 通用公共许可证 v3.0 (AGPLv3)</strong>。</li>
-          <li><strong>超过10人的企业/组织:</strong> <strong>必须</strong> 获取 <strong>商业许可证 (Commercial License)</strong>。</li>
-        </ul>
-      </section>
-
-      <section class="mb-8">
-        <h2 class="text-xl font-semibold mb-4 text-gray-900">定义："10人及以下"</h2>
-        <p class="text-gray-700">
-          指在您的组织（包括公司、非营利组织、政府机构、教育机构等任何实体）中，能够访问、使用或以任何方式直接或间接受益于本软件（Cherry
-          Studio）功能的个人总数不超过10人。这包括但不限于开发者、测试人员、运营人员、最终用户、通过集成系统间接使用者等。
         </p>
-      </section>
 
-      <section class="mb-8">
+        <section class="mb-8">
-        <h2 class="text-xl font-semibold mb-4 text-gray-900">1. 开源许可证 (Open Source License): AGPLv3 - 适用于个人及10人及以下组织
+          <h2 class="mb-4 text-xl font-semibold text-gray-900">核心原则</h2>
-        </h2>
+          <ul class="list-disc space-y-2 pl-6 text-gray-700">
-        <ul class="list-disc pl-6 space-y-2 text-gray-700">
+            <li>
-          <li>如果您是个人用户，或者您的组织满足上述"10人及以下"的定义，您可以在 <strong>AGPLv3</strong> 的条款下自由使用、修改和分发 Cherry Studio。AGPLv3 的完整文本可以访问
+              <strong>个人用户 和 10人及以下企业/组织:</strong> 默认适用
-            <a href="https://www.gnu.org/licenses/agpl-3.0.html"
+              <strong>GNU Affero 通用公共许可证 v3.0 (AGPLv3)</strong>。
-              class="text-blue-600 hover:underline">https://www.gnu.org/licenses/agpl-3.0.html</a> 获取。
+            </li>
-          </li>
+            <li>
-          <li><strong>核心义务：</strong> AGPLv3 的一个关键要求是，如果您修改了 Cherry Studio 并通过网络提供服务，或者分发了修改后的版本，您必须以 AGPLv3
+              <strong>超过10人的企业/组织:</strong> <strong>必须</strong> 获取
-            许可证向接收者提供相应的<strong>完整源代码</strong>。即使您符合"10人及以下"的标准，如果您希望避免此源代码公开义务，您也需要考虑获取商业许可证（见下文）。</li>
+              <strong>商业许可证 (Commercial License)</strong>。
-          <li>使用前请务必仔细阅读并理解 AGPLv3 的所有条款。</li>
+            </li>
-        </ul>
+          </ul>
-      </section>
+        </section>
 
-      <section class="mb-8">
+        <section class="mb-8">
-        <h2 class="text-xl font-semibold mb-4 text-gray-900">2. 商业许可证 (Commercial License) - 适用于超过10人的组织，或希望规避 AGPLv3
+          <h2 class="mb-4 text-xl font-semibold text-gray-900">定义："10人及以下"</h2>
-          义务的用户</h2>
+          <p class="text-gray-700">
-        <ul class="list-disc pl-6 space-y-2 text-gray-700">
+            指在您的组织（包括公司、非营利组织、政府机构、教育机构等任何实体）中，能够访问、使用或以任何方式直接或间接受益于本软件（Cherry
-          <li><strong>强制要求：</strong>
+            Studio）功能的个人总数不超过10人。这包括但不限于开发者、测试人员、运营人员、最终用户、通过集成系统间接使用者等。
-            如果您的组织<strong>不</strong>满足上述"10人及以下"的定义（即有11人或更多人可以访问、使用或受益于本软件），您<strong>必须</strong>联系我们获取并签署一份商业许可证才能使用
+          </p>
-            Cherry Studio。</li>
+        </section>
-          <li><strong>自愿选择：</strong> 即使您的组织满足"10人及以下"的条件，但如果您的使用场景<strong>无法满足 AGPLv3
-              的条款要求</strong>（特别是关于<strong>源代码公开</strong>的义务），或者您需要 AGPLv3 <strong>未提供</strong>的特定商业条款（如保证、赔偿、无 Copyleft
-            限制等），您也<strong>必须</strong>联系我们获取并签署一份商业许可证。</li>
-          <li><strong>需要商业许可证的常见情况包括（但不限于）：</strong>
-            <ul class="list-disc pl-6 mt-2 space-y-1">
-              <li>您的组织规模超过10人。</li>
-              <li>（无论组织规模）您希望分发修改过的 Cherry Studio 版本，但<strong>不希望</strong>根据 AGPLv3 公开您修改部分的源代码。</li>
-              <li>（无论组织规模）您希望基于修改过的 Cherry Studio 提供网络服务（SaaS），但<strong>不希望</strong>根据 AGPLv3 向服务使用者提供修改后的源代码。</li>
-              <li>（无论组织规模）您的公司政策、客户合同或项目要求不允许使用 AGPLv3 许可的软件，或要求闭源分发及保密。</li>
-            </ul>
-          </li>
-          <li><strong>获取商业许可：</strong> 请通过邮箱 <a href="mailto:bd@cherry-ai.com"
-              class="text-blue-600 hover:underline">bd@cherry-ai.com</a> 联系 Cherry Studio 开发团队洽谈商业授权事宜。</li>
-        </ul>
-      </section>
 
-      <section class="mb-8">
+        <section class="mb-8">
-        <h2 class="text-xl font-semibold mb-4 text-gray-900">3. 贡献 (Contributions)</h2>
+          <h2 class="mb-4 text-xl font-semibold text-gray-900">
-        <ul class="list-disc pl-6 space-y-2 text-gray-700">
+            1. 开源许可证 (Open Source License): AGPLv3 - 适用于个人及10人及以下组织
-          <li>我们欢迎社区对 Cherry Studio 的贡献。所有向本项目提交的贡献都将被视为在 <strong>AGPLv3</strong> 许可证下提供。</li>
+          </h2>
-          <li>通过向本项目提交贡献（例如通过 Pull Request），即表示您同意您的代码以 AGPLv3 许可证授权给本项目及所有后续使用者（无论这些使用者最终遵循 AGPLv3 还是商业许可）。</li>
+          <ul class="list-disc space-y-2 pl-6 text-gray-700">
-          <li>您也理解并同意，您的贡献可能会被包含在根据商业许可证分发的 Cherry Studio 版本中。</li>
+            <li>
-        </ul>
+              如果您是个人用户，或者您的组织满足上述"10人及以下"的定义，您可以在
-      </section>
+              <strong>AGPLv3</strong> 的条款下自由使用、修改和分发 Cherry Studio。AGPLv3 的完整文本可以访问
+              <a href="https://www.gnu.org/licenses/agpl-3.0.html" class="text-blue-600 hover:underline"
+                >https://www.gnu.org/licenses/agpl-3.0.html</a
+              >
+              获取。
+            </li>
+            <li>
+              <strong>核心义务：</strong> AGPLv3 的一个关键要求是，如果您修改了 Cherry Studio
+              并通过网络提供服务，或者分发了修改后的版本，您必须以 AGPLv3
+              许可证向接收者提供相应的<strong>完整源代码</strong>。即使您符合"10人及以下"的标准，如果您希望避免此源代码公开义务，您也需要考虑获取商业许可证（见下文）。
+            </li>
+            <li>使用前请务必仔细阅读并理解 AGPLv3 的所有条款。</li>
+          </ul>
+        </section>
 
-      <section class="mb-8">
+        <section class="mb-8">
-        <h2 class="text-xl font-semibold mb-4 text-gray-900">4. 其他条款 (Other Terms)</h2>
+          <h2 class="mb-4 text-xl font-semibold text-gray-900">
-        <ul class="list-disc pl-6 space-y-2 text-gray-700">
+            2. 商业许可证 (Commercial License) - 适用于超过10人的组织，或希望规避 AGPLv3 义务的用户
-          <li>关于商业许可证的具体条款和条件，以双方签署的正式商业许可协议为准。</li>
+          </h2>
-          <li>项目维护者保留根据需要更新本许可政策（包括用户规模定义和阈值）的权利。相关更新将通过项目官方渠道（如代码仓库、官方网站）进行通知。</li>
+          <ul class="list-disc space-y-2 pl-6 text-gray-700">
-        </ul>
+            <li>
-      </section>
+              <strong>强制要求：</strong>
+              如果您的组织<strong>不</strong>满足上述"10人及以下"的定义（即有11人或更多人可以访问、使用或受益于本软件），您<strong>必须</strong>联系我们获取并签署一份商业许可证才能使用
+              Cherry Studio。
+            </li>
+            <li>
+              <strong>自愿选择：</strong> 即使您的组织满足"10人及以下"的条件，但如果您的使用场景<strong
+                >无法满足 AGPLv3 的条款要求</strong
+              >（特别是关于<strong>源代码公开</strong>的义务），或者您需要 AGPLv3
+              <strong>未提供</strong>的特定商业条款（如保证、赔偿、无 Copyleft
+              限制等），您也<strong>必须</strong>联系我们获取并签署一份商业许可证。
+            </li>
+            <li>
+              <strong>需要商业许可证的常见情况包括（但不限于）：</strong>
+              <ul class="mt-2 list-disc space-y-1 pl-6">
+                <li>您的组织规模超过10人。</li>
+                <li>
+                  （无论组织规模）您希望分发修改过的 Cherry Studio 版本，但<strong>不希望</strong>根据 AGPLv3
+                  公开您修改部分的源代码。
+                </li>
+                <li>
+                  （无论组织规模）您希望基于修改过的 Cherry Studio 提供网络服务（SaaS），但<strong>不希望</strong>根据
+                  AGPLv3 向服务使用者提供修改后的源代码。
+                </li>
+                <li>
+                  （无论组织规模）您的公司政策、客户合同或项目要求不允许使用 AGPLv3 许可的软件，或要求闭源分发及保密。
+                </li>
+              </ul>
+            </li>
+            <li>
+              <strong>获取商业许可：</strong> 请通过邮箱
+              <a href="mailto:bd@cherry-ai.com" class="text-blue-600 hover:underline">bd@cherry-ai.com</a> 联系 Cherry
+              Studio 开发团队洽谈商业授权事宜。
+            </li>
+          </ul>
+        </section>
+
+        <section class="mb-8">
+          <h2 class="mb-4 text-xl font-semibold text-gray-900">3. 贡献 (Contributions)</h2>
+          <ul class="list-disc space-y-2 pl-6 text-gray-700">
+            <li>
+              我们欢迎社区对 Cherry Studio 的贡献。所有向本项目提交的贡献都将被视为在
+              <strong>AGPLv3</strong> 许可证下提供。
+            </li>
+            <li>
+              通过向本项目提交贡献（例如通过 Pull Request），即表示您同意您的代码以 AGPLv3
+              许可证授权给本项目及所有后续使用者（无论这些使用者最终遵循 AGPLv3 还是商业许可）。
+            </li>
+            <li>您也理解并同意，您的贡献可能会被包含在根据商业许可证分发的 Cherry Studio 版本中。</li>
+          </ul>
+        </section>
+
+        <section class="mb-8">
+          <h2 class="mb-4 text-xl font-semibold text-gray-900">4. 其他条款 (Other Terms)</h2>
+          <ul class="list-disc space-y-2 pl-6 text-gray-700">
+            <li>关于商业许可证的具体条款和条件，以双方签署的正式商业许可协议为准。</li>
+            <li>
+              项目维护者保留根据需要更新本许可政策（包括用户规模定义和阈值）的权利。相关更新将通过项目官方渠道（如代码仓库、官方网站）进行通知。
+            </li>
+          </ul>
+        </section>
+      </div>
+
+      <hr class="my-12 border-gray-300" />
+
+      <!-- English Version -->
+      <div>
+        <h1 class="mb-8 text-3xl font-bold text-gray-900">Licensing</h1>
+
+        <p class="mb-6 text-gray-700">This project employs a <strong>User-Segmented Dual Licensing</strong> model.</p>
+
+        <section class="mb-8">
+          <h2 class="mb-4 text-xl font-semibold text-gray-900">Core Principle</h2>
+          <ul class="list-disc space-y-2 pl-6 text-gray-700">
+            <li>
+              <strong>Individual Users and Organizations with 10 or Fewer Individuals:</strong> Governed by default
+              under the <strong>GNU Affero General Public License v3.0 (AGPLv3)</strong>.
+            </li>
+            <li>
+              <strong>Organizations with More Than 10 Individuals:</strong> <strong>Must</strong> obtain a
+              <strong>Commercial License</strong>.
+            </li>
+          </ul>
+        </section>
+
+        <section class="mb-8">
+          <h2 class="mb-4 text-xl font-semibold text-gray-900">Definition: "10 or Fewer Individuals"</h2>
+          <p class="text-gray-700">
+            Refers to any organization (including companies, non-profits, government agencies, educational institutions,
+            etc.) where the total number of individuals who can access, use, or in any way directly or indirectly
+            benefit from the functionality of this software (Cherry Studio) does not exceed 10. This includes, but is
+            not limited to, developers, testers, operations staff, end-users, and indirect users via integrated systems.
+          </p>
+        </section>
+
+        <section class="mb-8">
+          <h2 class="mb-4 text-xl font-semibold text-gray-900">
+            1. Open Source License: AGPLv3 - For Individuals and Organizations of 10 or Fewer
+          </h2>
+          <ul class="list-disc space-y-2 pl-6 text-gray-700">
+            <li>
+              If you are an individual user, or if your organization meets the "10 or Fewer Individuals" definition
+              above, you are free to use, modify, and distribute Cherry Studio under the terms of the
+              <strong>AGPLv3</strong>. The full text of the AGPLv3 can be found at
+              <a href="https://www.gnu.org/licenses/agpl-3.0.html" class="text-blue-600 hover:underline"
+                >https://www.gnu.org/licenses/agpl-3.0.html</a
+              >.
+            </li>
+            <li>
+              <strong>Core Obligation:</strong> A key requirement of the AGPLv3 is that if you modify Cherry Studio and
+              make it available over a network, or distribute the modified version, you must provide the
+              <strong>complete corresponding source code</strong> under the AGPLv3 license to the recipients. Even if
+              you qualify under the "10 or Fewer Individuals" rule, if you wish to avoid this source code disclosure
+              obligation, you will need to obtain a Commercial License (see below).
+            </li>
+            <li>Please read and understand the full terms of the AGPLv3 carefully before use.</li>
+          </ul>
+        </section>
+
+        <section class="mb-8">
+          <h2 class="mb-4 text-xl font-semibold text-gray-900">
+            2. Commercial License - For Organizations with More Than 10 Individuals, or Users Needing to Avoid AGPLv3
+            Obligations
+          </h2>
+          <ul class="list-disc space-y-2 pl-6 text-gray-700">
+            <li>
+              <strong>Mandatory Requirement:</strong> If your organization does <strong>not</strong> meet the "10 or
+              Fewer Individuals" definition above (i.e., 11 or more individuals can access, use, or benefit from the
+              software), you <strong>must</strong> contact us to obtain and execute a Commercial License to use Cherry
+              Studio.
+            </li>
+            <li>
+              <strong>Voluntary Option:</strong> Even if your organization meets the "10 or Fewer Individuals"
+              condition, if your intended use case
+              <strong>cannot comply with the terms of the AGPLv3</strong> (particularly the obligations regarding
+              <strong>source code disclosure</strong>), or if you require specific commercial terms
+              <strong>not offered</strong> by the AGPLv3 (such as warranties, indemnities, or freedom from copyleft
+              restrictions), you also <strong>must</strong> contact us to obtain and execute a Commercial License.
+            </li>
+            <li>
+              <strong>Common scenarios requiring a Commercial License include (but are not limited to):</strong>
+              <ul class="mt-2 list-disc space-y-1 pl-6">
+                <li>
+                  Your organization has more than 10 individuals who can access, use, or benefit from the software.
+                </li>
+                <li>
+                  (Regardless of organization size) You wish to distribute a modified version of Cherry Studio but
+                  <strong>do not want</strong> to disclose the source code of your modifications under AGPLv3.
+                </li>
+                <li>
+                  (Regardless of organization size) You wish to provide a network service (SaaS) based on a modified
+                  version of Cherry Studio but <strong>do not want</strong> to provide the modified source code to users
+                  of the service under AGPLv3.
+                </li>
+                <li>
+                  (Regardless of organization size) Your corporate policies, client contracts, or project requirements
+                  prohibit the use of AGPLv3-licensed software or mandate closed-source distribution and
+                  confidentiality.
+                </li>
+              </ul>
+            </li>
+            <li>
+              <strong>Obtaining a Commercial License:</strong> Please contact the Cherry Studio development team via
+              email at <a href="mailto:bd@cherry-ai.com" class="text-blue-600 hover:underline">bd@cherry-ai.com</a> to
+              discuss commercial licensing options.
+            </li>
+          </ul>
+        </section>
+
+        <section class="mb-8">
+          <h2 class="mb-4 text-xl font-semibold text-gray-900">3. Contributions</h2>
+          <ul class="list-disc space-y-2 pl-6 text-gray-700">
+            <li>
+              We welcome community contributions to Cherry Studio. All contributions submitted to this project are
+              considered to be offered under the <strong>AGPLv3</strong> license.
+            </li>
+            <li>
+              By submitting a contribution to this project (e.g., via a Pull Request), you agree to license your code
+              under the AGPLv3 to the project and all its downstream users (regardless of whether those users ultimately
+              operate under AGPLv3 or a Commercial License).
+            </li>
+            <li>
+              You also understand and agree that your contribution may be included in distributions of Cherry Studio
+              offered under our commercial license.
+            </li>
+          </ul>
+        </section>
+
+        <section class="mb-8">
+          <h2 class="mb-4 text-xl font-semibold text-gray-900">4. Other Terms</h2>
+          <ul class="list-disc space-y-2 pl-6 text-gray-700">
+            <li>
+              The specific terms and conditions of the Commercial License are governed by the formal commercial license
+              agreement signed by both parties.
+            </li>
+            <li>
+              The project maintainers reserve the right to update this licensing policy (including the definition and
+              threshold for user count) as needed. Updates will be communicated through official project channels (e.g.,
+              code repository, official website).
+            </li>
+          </ul>
+        </section>
+      </div>
     </div>
-
+  </body>
-    <hr class="my-12 border-gray-300">
+</html>
-
-    <!-- English Version -->
-    <div>
-      <h1 class="text-3xl font-bold mb-8 text-gray-900">Licensing</h1>
-
-      <p class="mb-6 text-gray-700">This project employs a <strong>User-Segmented Dual Licensing</strong> model.</p>
-
-      <section class="mb-8">
-        <h2 class="text-xl font-semibold mb-4 text-gray-900">Core Principle</h2>
-        <ul class="list-disc pl-6 space-y-2 text-gray-700">
-          <li><strong>Individual Users and Organizations with 10 or Fewer Individuals:</strong> Governed by default
-            under the <strong>GNU Affero General Public License v3.0 (AGPLv3)</strong>.</li>
-          <li><strong>Organizations with More Than 10 Individuals:</strong> <strong>Must</strong> obtain a
-            <strong>Commercial License</strong>.
-          </li>
-        </ul>
-      </section>
-
-      <section class="mb-8">
-        <h2 class="text-xl font-semibold mb-4 text-gray-900">Definition: "10 or Fewer Individuals"</h2>
-        <p class="text-gray-700">
-          Refers to any organization (including companies, non-profits, government agencies, educational institutions,
-          etc.) where the total number of individuals who can access, use, or in any way directly or indirectly benefit
-          from the functionality of this software (Cherry Studio) does not exceed 10. This includes, but is not limited
-          to, developers, testers, operations staff, end-users, and indirect users via integrated systems.
-        </p>
-      </section>
-
-      <section class="mb-8">
-        <h2 class="text-xl font-semibold mb-4 text-gray-900">1. Open Source License: AGPLv3 - For Individuals and
-          Organizations of 10 or Fewer</h2>
-        <ul class="list-disc pl-6 space-y-2 text-gray-700">
-          <li>If you are an individual user, or if your organization meets the "10 or Fewer Individuals" definition
-            above, you are free to use, modify, and distribute Cherry Studio under the terms of the
-            <strong>AGPLv3</strong>. The full text of the AGPLv3 can be found at <a
-              href="https://www.gnu.org/licenses/agpl-3.0.html"
-              class="text-blue-600 hover:underline">https://www.gnu.org/licenses/agpl-3.0.html</a>.
-          </li>
-          <li><strong>Core Obligation:</strong> A key requirement of the AGPLv3 is that if you modify Cherry Studio and
-            make it available over a network, or distribute the modified version, you must provide the <strong>complete
-              corresponding source code</strong> under the AGPLv3 license to the recipients. Even if you qualify under
-            the "10 or Fewer Individuals" rule, if you wish to avoid this source code disclosure obligation, you will
-            need to obtain a Commercial License (see below).</li>
-          <li>Please read and understand the full terms of the AGPLv3 carefully before use.</li>
-        </ul>
-      </section>
-
-      <section class="mb-8">
-        <h2 class="text-xl font-semibold mb-4 text-gray-900">2. Commercial License - For Organizations with More Than 10
-          Individuals, or Users Needing to Avoid AGPLv3 Obligations</h2>
-        <ul class="list-disc pl-6 space-y-2 text-gray-700">
-          <li><strong>Mandatory Requirement:</strong> If your organization does <strong>not</strong> meet the "10 or
-            Fewer Individuals" definition above (i.e., 11 or more individuals can access, use, or benefit from the
-            software), you <strong>must</strong> contact us to obtain and execute a Commercial License to use Cherry
-            Studio.</li>
-          <li><strong>Voluntary Option:</strong> Even if your organization meets the "10 or Fewer Individuals"
-            condition, if your intended use case <strong>cannot comply with the terms of the AGPLv3</strong>
-            (particularly the obligations regarding <strong>source code disclosure</strong>), or if you require specific
-            commercial terms <strong>not offered</strong> by the AGPLv3 (such as warranties, indemnities, or freedom
-            from copyleft restrictions), you also <strong>must</strong> contact us to obtain and execute a Commercial
-            License.</li>
-          <li><strong>Common scenarios requiring a Commercial License include (but are not limited to):</strong>
-            <ul class="list-disc pl-6 mt-2 space-y-1">
-              <li>Your organization has more than 10 individuals who can access, use, or benefit from the software.</li>
-              <li>(Regardless of organization size) You wish to distribute a modified version of Cherry Studio but
-                <strong>do not want</strong> to disclose the source code of your modifications under AGPLv3.
-              </li>
-              <li>(Regardless of organization size) You wish to provide a network service (SaaS) based on a modified
-                version of Cherry Studio but <strong>do not want</strong> to provide the modified source code to users
-                of the service under AGPLv3.</li>
-              <li>(Regardless of organization size) Your corporate policies, client contracts, or project requirements
-                prohibit the use of AGPLv3-licensed software or mandate closed-source distribution and confidentiality.
-              </li>
-            </ul>
-          </li>
-          <li><strong>Obtaining a Commercial License:</strong> Please contact the Cherry Studio development team via
-            email at <a href="mailto:bd@cherry-ai.com" class="text-blue-600 hover:underline">bd@cherry-ai.com</a> to
-            discuss commercial licensing options.</li>
-        </ul>
-      </section>
-
-      <section class="mb-8">
-        <h2 class="text-xl font-semibold mb-4 text-gray-900">3. Contributions</h2>
-        <ul class="list-disc pl-6 space-y-2 text-gray-700">
-          <li>We welcome community contributions to Cherry Studio. All contributions submitted to this project are
-            considered to be offered under the <strong>AGPLv3</strong> license.</li>
-          <li>By submitting a contribution to this project (e.g., via a Pull Request), you agree to license your code
-            under the AGPLv3 to the project and all its downstream users (regardless of whether those users ultimately
-            operate under AGPLv3 or a Commercial License).</li>
-          <li>You also understand and agree that your contribution may be included in distributions of Cherry Studio
-            offered under our commercial license.</li>
-        </ul>
-      </section>
-
-      <section class="mb-8">
-        <h2 class="text-xl font-semibold mb-4 text-gray-900">4. Other Terms</h2>
-        <ul class="list-disc pl-6 space-y-2 text-gray-700">
-          <li>The specific terms and conditions of the Commercial License are governed by the formal commercial license
-            agreement signed by both parties.</li>
-          <li>The project maintainers reserve the right to update this licensing policy (including the definition and
-            threshold for user count) as needed. Updates will be communicated through official project channels (e.g.,
-            code repository, official website).</li>
-        </ul>
-      </section>
-    </div>
-  </div>
-</body>
-
-</html>

resources/cherry-studio/releases.html

@@ -12,18 +12,18 @@
 
   <body id="app">
     <div :class="isDark ? 'dark-bg' : 'bg'" class="min-h-screen">
-      <div class="max-w-3xl mx-auto py-12 px-4">
+      <div class="mx-auto max-w-3xl px-4 py-12">
-        <h1 class="text-3xl font-bold mb-8" :class="isDark ? 'text-white' : 'text-gray-900'">Release Timeline</h1>
+        <h1 class="mb-8 text-3xl font-bold" :class="isDark ? 'text-white' : 'text-gray-900'">Release Timeline</h1>
 
         <!-- Loading状态 -->
-        <div v-if="loading" class="text-center py-8">
+        <div v-if="loading" class="py-8 text-center">
           <div
-            class="inline-block animate-spin rounded-full h-8 w-8 border-4"
+            class="inline-block h-8 w-8 animate-spin rounded-full border-4"
             :class="isDark ? 'border-gray-700 border-t-blue-500' : 'border-gray-300 border-t-blue-500'"></div>
         </div>
 
         <!-- Error 状态 -->
-        <div v-else-if="error" class="text-red-500 text-center py-8">{{ error }}</div>
+        <div v-else-if="error" class="py-8 text-center text-red-500">{{ error }}</div>
 
         <!-- Release 列表 -->
         <div v-else class="space-y-8">
@@ -32,21 +32,21 @@
             :key="release.id"
             class="relative pl-8"
             :class="isDark ? 'border-l-2 border-gray-700' : 'border-l-2 border-gray-200'">
-            <div class="absolute -left-2 top-0 w-4 h-4 rounded-full bg-green-500"></div>
+            <div class="absolute top-0 -left-2 h-4 w-4 rounded-full bg-green-500"></div>
             <div
-              class="rounded-lg shadow-sm p-6 transition-shadow"
+              class="rounded-lg p-6 shadow-sm transition-shadow"
               :class="isDark ? 'bg-black hover:shadow-md hover:shadow-black' : 'bg-white hover:shadow-md'">
-              <div class="flex items-start justify-between mb-4">
+              <div class="mb-4 flex items-start justify-between">
                 <div>
                   <h2 class="text-xl font-semibold" :class="isDark ? 'text-white' : 'text-gray-900'">
                     {{ release.name || release.tag_name }}
                   </h2>
-                  <p class="text-sm mt-1" :class="isDark ? 'text-gray-400' : 'text-gray-500'">
+                  <p class="mt-1 text-sm" :class="isDark ? 'text-gray-400' : 'text-gray-500'">
                     {{ formatDate(release.published_at) }}
                   </p>
                 </div>
                 <span
-                  class="inline-flex items-center px-3 py-1 rounded-full text-sm font-medium"
+                  class="inline-flex items-center rounded-full px-3 py-1 text-sm font-medium"
                   :class="isDark ? 'bg-green-900 text-green-200' : 'bg-green-100 text-green-800'">
                   {{ release.tag_name }}
                 </span>

resources/database/drizzle/0000_confused_wendigo.sql

@@ -0,0 +1,53 @@
+--> statement-breakpoint
+CREATE TABLE `migrations` (
+	`version` integer PRIMARY KEY NOT NULL,
+	`tag` text NOT NULL,
+	`executed_at` integer NOT NULL
+);
+
+CREATE TABLE `agents` (
+	`id` text PRIMARY KEY NOT NULL,
+	`type` text NOT NULL,
+	`name` text NOT NULL,
+	`description` text,
+	`accessible_paths` text,
+	`instructions` text,
+	`model` text NOT NULL,
+	`plan_model` text,
+	`small_model` text,
+	`mcps` text,
+	`allowed_tools` text,
+	`configuration` text,
+	`created_at` text NOT NULL,
+	`updated_at` text NOT NULL
+);
+
+--> statement-breakpoint
+CREATE TABLE `sessions` (
+	`id` text PRIMARY KEY NOT NULL,
+	`agent_type` text NOT NULL,
+	`agent_id` text NOT NULL,
+	`name` text NOT NULL,
+	`description` text,
+	`accessible_paths` text,
+	`instructions` text,
+	`model` text NOT NULL,
+	`plan_model` text,
+	`small_model` text,
+	`mcps` text,
+	`allowed_tools` text,
+	`configuration` text,
+	`created_at` text NOT NULL,
+	`updated_at` text NOT NULL
+);
+
+--> statement-breakpoint
+CREATE TABLE `session_messages` (
+	`id` integer PRIMARY KEY AUTOINCREMENT NOT NULL,
+	`session_id` text NOT NULL,
+	`role` text NOT NULL,
+	`content` text NOT NULL,
+	`metadata` text,
+	`created_at` text NOT NULL,
+	`updated_at` text NOT NULL
+);

resources/database/drizzle/0001_woozy_captain_flint.sql

@@ -0,0 +1 @@
+ALTER TABLE `session_messages` ADD `agent_session_id` text DEFAULT '';

resources/database/drizzle/meta/0000_snapshot.json

@@ -0,0 +1,331 @@
+{
+  "version": "6",
+  "dialect": "sqlite",
+  "id": "35efb412-0230-4767-9c76-7b7c4d40369f",
+  "prevId": "00000000-0000-0000-0000-000000000000",
+  "tables": {
+    "agents": {
+      "name": "agents",
+      "columns": {
+        "id": {
+          "name": "id",
+          "type": "text",
+          "primaryKey": true,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "type": {
+          "name": "type",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "name": {
+          "name": "name",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "description": {
+          "name": "description",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "accessible_paths": {
+          "name": "accessible_paths",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "instructions": {
+          "name": "instructions",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "model": {
+          "name": "model",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "plan_model": {
+          "name": "plan_model",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "small_model": {
+          "name": "small_model",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "mcps": {
+          "name": "mcps",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "allowed_tools": {
+          "name": "allowed_tools",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "configuration": {
+          "name": "configuration",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "created_at": {
+          "name": "created_at",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "updated_at": {
+          "name": "updated_at",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        }
+      },
+      "indexes": {},
+      "foreignKeys": {},
+      "compositePrimaryKeys": {},
+      "uniqueConstraints": {},
+      "checkConstraints": {}
+    },
+    "session_messages": {
+      "name": "session_messages",
+      "columns": {
+        "id": {
+          "name": "id",
+          "type": "integer",
+          "primaryKey": true,
+          "notNull": true,
+          "autoincrement": true
+        },
+        "session_id": {
+          "name": "session_id",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "role": {
+          "name": "role",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "content": {
+          "name": "content",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "metadata": {
+          "name": "metadata",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "created_at": {
+          "name": "created_at",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "updated_at": {
+          "name": "updated_at",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        }
+      },
+      "indexes": {},
+      "foreignKeys": {},
+      "compositePrimaryKeys": {},
+      "uniqueConstraints": {},
+      "checkConstraints": {}
+    },
+    "migrations": {
+      "name": "migrations",
+      "columns": {
+        "version": {
+          "name": "version",
+          "type": "integer",
+          "primaryKey": true,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "tag": {
+          "name": "tag",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "executed_at": {
+          "name": "executed_at",
+          "type": "integer",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        }
+      },
+      "indexes": {},
+      "foreignKeys": {},
+      "compositePrimaryKeys": {},
+      "uniqueConstraints": {},
+      "checkConstraints": {}
+    },
+    "sessions": {
+      "name": "sessions",
+      "columns": {
+        "id": {
+          "name": "id",
+          "type": "text",
+          "primaryKey": true,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "agent_type": {
+          "name": "agent_type",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "agent_id": {
+          "name": "agent_id",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "name": {
+          "name": "name",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "description": {
+          "name": "description",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "accessible_paths": {
+          "name": "accessible_paths",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "instructions": {
+          "name": "instructions",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "model": {
+          "name": "model",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "plan_model": {
+          "name": "plan_model",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "small_model": {
+          "name": "small_model",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "mcps": {
+          "name": "mcps",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "allowed_tools": {
+          "name": "allowed_tools",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "configuration": {
+          "name": "configuration",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "created_at": {
+          "name": "created_at",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "updated_at": {
+          "name": "updated_at",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        }
+      },
+      "indexes": {},
+      "foreignKeys": {},
+      "compositePrimaryKeys": {},
+      "uniqueConstraints": {},
+      "checkConstraints": {}
+    }
+  },
+  "views": {},
+  "enums": {},
+  "_meta": {
+    "schemas": {},
+    "tables": {},
+    "columns": {}
+  },
+  "internal": {
+    "indexes": {}
+  }
+}

resources/database/drizzle/meta/0001_snapshot.json

@@ -0,0 +1,339 @@
+{
+  "version": "6",
+  "dialect": "sqlite",
+  "id": "dabab6db-a2cd-4e96-b06e-6cb87d445a87",
+  "prevId": "35efb412-0230-4767-9c76-7b7c4d40369f",
+  "tables": {
+    "agents": {
+      "name": "agents",
+      "columns": {
+        "id": {
+          "name": "id",
+          "type": "text",
+          "primaryKey": true,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "type": {
+          "name": "type",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "name": {
+          "name": "name",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "description": {
+          "name": "description",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "accessible_paths": {
+          "name": "accessible_paths",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "instructions": {
+          "name": "instructions",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "model": {
+          "name": "model",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "plan_model": {
+          "name": "plan_model",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "small_model": {
+          "name": "small_model",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "mcps": {
+          "name": "mcps",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "allowed_tools": {
+          "name": "allowed_tools",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "configuration": {
+          "name": "configuration",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "created_at": {
+          "name": "created_at",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "updated_at": {
+          "name": "updated_at",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        }
+      },
+      "indexes": {},
+      "foreignKeys": {},
+      "compositePrimaryKeys": {},
+      "uniqueConstraints": {},
+      "checkConstraints": {}
+    },
+    "session_messages": {
+      "name": "session_messages",
+      "columns": {
+        "id": {
+          "name": "id",
+          "type": "integer",
+          "primaryKey": true,
+          "notNull": true,
+          "autoincrement": true
+        },
+        "session_id": {
+          "name": "session_id",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "role": {
+          "name": "role",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "content": {
+          "name": "content",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "agent_session_id": {
+          "name": "agent_session_id",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false,
+          "default": "''"
+        },
+        "metadata": {
+          "name": "metadata",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "created_at": {
+          "name": "created_at",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "updated_at": {
+          "name": "updated_at",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        }
+      },
+      "indexes": {},
+      "foreignKeys": {},
+      "compositePrimaryKeys": {},
+      "uniqueConstraints": {},
+      "checkConstraints": {}
+    },
+    "migrations": {
+      "name": "migrations",
+      "columns": {
+        "version": {
+          "name": "version",
+          "type": "integer",
+          "primaryKey": true,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "tag": {
+          "name": "tag",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "executed_at": {
+          "name": "executed_at",
+          "type": "integer",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        }
+      },
+      "indexes": {},
+      "foreignKeys": {},
+      "compositePrimaryKeys": {},
+      "uniqueConstraints": {},
+      "checkConstraints": {}
+    },
+    "sessions": {
+      "name": "sessions",
+      "columns": {
+        "id": {
+          "name": "id",
+          "type": "text",
+          "primaryKey": true,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "agent_type": {
+          "name": "agent_type",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "agent_id": {
+          "name": "agent_id",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "name": {
+          "name": "name",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "description": {
+          "name": "description",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "accessible_paths": {
+          "name": "accessible_paths",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "instructions": {
+          "name": "instructions",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "model": {
+          "name": "model",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "plan_model": {
+          "name": "plan_model",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "small_model": {
+          "name": "small_model",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "mcps": {
+          "name": "mcps",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "allowed_tools": {
+          "name": "allowed_tools",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "configuration": {
+          "name": "configuration",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": false,
+          "autoincrement": false
+        },
+        "created_at": {
+          "name": "created_at",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        },
+        "updated_at": {
+          "name": "updated_at",
+          "type": "text",
+          "primaryKey": false,
+          "notNull": true,
+          "autoincrement": false
+        }
+      },
+      "indexes": {},
+      "foreignKeys": {},
+      "compositePrimaryKeys": {},
+      "uniqueConstraints": {},
+      "checkConstraints": {}
+    }
+  },
+  "views": {},
+  "enums": {},
+  "_meta": {
+    "schemas": {},
+    "tables": {},
+    "columns": {}
+  },
+  "internal": {
+    "indexes": {}
+  }
+}