* feat: replace update dialog handling with quit and install functionality
* refactor: remove App_ShowUpdateDialog and implement App_QuitAndInstall in IpcChannel
* update ipc.ts to handle quit and install action
* modify AppUpdater to include quitAndInstall method
* adjust preload index to invoke new quit and install action
* enhance AboutSettings to manage update dialog state and trigger quit and install
* fix(AboutSettings): handle null update info in update dialog state management
* fix(UpdateDialog): improve error handling during update installation and enhance release notes processing
* fix(AppUpdater): remove redundant assignment of releaseInfo after update download
* fix(IpcChannel): remove UpdateDownloadedCancelled enum value
* format code
* fix(UpdateDialog): enhance installation process with loading state and error handling
* update i18n
* fix(i18n): Auto update translations for PR #10569
* feat(UpdateAppButton): integrate UpdateDialog and update button functionality for better user experience
* fix(UpdateDialog): update installation handler to support async operation and ensure modal closes after installation
* refactor(AppUpdater.test): remove deprecated formatReleaseNotes tests to streamline test suite
* refactor(update-dialog): simplify dialog close handling
Replace onOpenChange with onClose prop to directly handle dialog closing
Remove redundant handleClose function and simplify button onPress handler
---------
Co-authored-by: GitHub Action <action@github.com>
Co-authored-by: icarus <eurfelux@gmail.com>
- Adjusted class names in Message component for better layout management.
- Modified margin in DropHintNode of NotesSidebar for improved spacing.
- Enhanced BackupService to remove 'notes_tree' from indexedDB during data restoration.
chore: bump version to 1.6.3 and add migration for missing providers #10425fix: #10425
- Updated the version from 158 to 159 in the persisted reducer configuration.
- Implemented a migration function to ensure missing system providers are added to the state during the migration to version 159, enhancing state consistency.
* feat(notes): add spell-check control
* feat(notes): add spell-check toggle to preview mode toolbar
* feat(settings): move spellcheck to global and use hook
* fix(ui): remove redundant scrollbar in side-by-side view
Changed GridContainer from styled(Scrollbar) to styled.div to
eliminate redundant horizontal scrollbar in multi-model horizontal
layout mode. The Scrollbar component is designed for vertical
scrolling and conflicts with horizontal layouts.
Fixes#10520
* fix(ui): restore vertical scrollbar for grid mode while preserving horizontal fix
Optimal solution: Use Scrollbar component as base to preserve auto-hide
behavior for vertical modes (grid, vertical, fold) while overriding its
overflow-y behavior for horizontal mode only.
This approach:
- Preserves the June 2025 UX optimization (auto-hide scrollbars)
- Fixes horizontal scrollbar issue from #10520
- Restores vertical scrolling for grid mode
- Maintains auto-hide behavior for all vertical scrolling modes
- Minimal change with no code duplication
The Scrollbar component provides scrollbar thumb auto-hide after 1.5s,
which enhances UX for vertical scrolling. By using CSS overrides only
for horizontal mode, we get the best of both worlds.
* chore: fix import sorting in MessageGroup.tsx
Unrelated to PR scope - fixing to unblock CI.
Auto-fixed via eslint --fix (moved Scrollbar import to correct position).
Also updated yarn.lock to resolve dependency sync.
* fix(ui): add explicit overflow declarations for all grid modes
Previous fix relied on CSS inheritance from Scrollbar base component,
but display: grid interferes with overflow property inheritance.
This iteration adds explicit overflow-y: auto and overflow-x: hidden
to grid, fold, vertical, and multi-select modes to ensure vertical
scrolling works reliably across all layouts.
- horizontal mode: overflow-y visible, overflow-x auto (unchanged)
- grid/fold/vertical modes: explicit overflow-y auto, overflow-x hidden
- multi-select mode: explicit overflow-y auto, overflow-x hidden
Fixes vertical scrollbar missing in grid mode reported by @EurFelux
* fix(Messages): adjust overflow behavior in message groups
Fix scrollbar issues by hiding vertical overflow in horizontal layout and simplifying overflow handling in grid layout
* feat(HorizontalScrollContainer): add classNames prop for container and content styling
allow custom styling of container and content via classNames prop
---------
Co-authored-by: icarus <eurfelux@gmail.com>
* chore: update electron dependency from 37.4.0 to 37.6.0
* feat(TopicsTab): add double click to edit topic name
Move double click handler from TopicName component to parent div to improve UX
* fix(TopicsTab): prevent topic edit on double click when already editing
* fix(TextFilePreview): make editor read-only but can be copied
* feat: add table auto-wrap feature for notes
* Revert "feat: add table auto-wrap feature for notes"
This reverts commit 7785f480b1.
* fix(reasoning): update deepseek model id regex pattern to match more variants
The previous regex pattern was too restrictive and didn't account for all possible deepseek model id formats. This change expands the pattern to support more variants while maintaining the same functionality.
* fix(reasoning): update deepseek model id regex pattern to match more variants
* fix(reasoning): improve regex pattern for deepseek model matching
Update the regex pattern to be more precise in matching deepseek model versions.
Add detailed comments explaining the pattern and note future improvements.
* feat(models): add GLM-4.6 model to supported list
Update model configuration to include new GLM-4.6 model and add it to the supported models for thinking token functionality
* feat(models): add claude sonnet 4.5 model to anthropic provider
* feat: implement auto-renaming feature for notes
* feat: motion effects for auto renaming in notes
* feat: add i18n for zh-tw for auto renaming in notes
* chore: lint
* feat: add GitHub Copilot CLI integration to coding tools
- Add githubCopilotCli to codeTools enum
- Support @github/copilot package installation
- Add 'copilot' executable command mapping
- Update Redux store to include GitHub Copilot CLI state
- Add GitHub Copilot CLI option to UI with proper provider mapping
- Implement environment variable handling for GitHub authentication
- Fix model selection logic to disable model choice for GitHub Copilot CLI
- Update launch validation to not require model selection for GitHub Copilot CLI
- Fix prepareLaunchEnvironment and executeLaunch to handle no-model scenario
This enables users to launch GitHub Copilot CLI directly from Cherry Studio's
code tools interface without needing to select a model, as GitHub Copilot CLI
uses GitHub's built-in models and authentication.
Signed-off-by: LeaderOnePro <leaderonepro@outlook.com>
* style: apply code formatting for GitHub Copilot CLI integration
Auto-fix code style inconsistencies using project's Biome formatter.
Resolves semicolon, comma, and quote style issues to match project standards.
Signed-off-by: LeaderOnePro <leaderonepro@outlook.com>
* feat: conditionally render model selector for GitHub Copilot CLI
- Hide model selector component when GitHub Copilot CLI is selected
- Maintain validation logic to allow GitHub Copilot CLI without model selection
- Improve UX by removing empty model dropdown for GitHub Copilot CLI
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
---------
Signed-off-by: LeaderOnePro <leaderonepro@outlook.com>
Co-authored-by: Claude <noreply@anthropic.com>
- Unify buildClaudeCodeSystemMessage implementation in shared package
- Refactor MessagesService to provide comprehensive message processing API
- Extract streaming logic, error handling, and header preparation into service methods
- Remove duplicate anthropic config from renderer, use shared implementation
- Update ClaudeCodeService to use append mode for custom instructions
- Improve type safety and request validation in message processing
- Replace @anthropic-ai/claude-code with @anthropic-ai/claude-agent-sdk@0.1.1
- Update all import statements across 4 files
- Migrate patch for Electron compatibility (fork vs spawn)
- Handle breaking changes: replace appendSystemPrompt with systemPrompt preset
- Add settingSources configuration for filesystem settings
- Update vendor path in build scripts
- Update package name mapping in CodeToolsService
* feat(models): add gpt5_codex model support
Add support for gpt5_codex model type in model configuration and type definitions. Update getThinkModelType to handle codex variant of gpt5 models.
* feat(models): add gpt-5-codex model logo and update logo mapping
Add new GPT-5-Codex model logo image and include it in the logo mapping configuration
Implement functionality to show files/folders in system explorer through IPC. Includes channel definition, preload API, main handler, and error handling for non-existent paths.
* add new provider: OVMS(openvino model server)
Signed-off-by: Ma, Kejiang <kj.ma@intel.com>
* remove useless comments
* add note: support windows only
* fix eslint error; add migrate for ovms provider
Signed-off-by: Ma, Kejiang <kj.ma@intel.com>
* fix ci error after rebase
Signed-off-by: Ma, Kejiang <kj.ma@intel.com>
* modifications base on reviewers' comments
Signed-off-by: Ma, Kejiang <kj.ma@intel.com>
* show intel-ovms provider only on windows and intel cpu
Signed-off-by: Ma, Kejiang <kj.ma@intel.com>
* complete i18n for intel ovms
Signed-off-by: Ma, Kejiang <kj.ma@intel.com>
* update ovms 2025.3; apply patch for model qwen3-8b on local
Signed-off-by: Ma, Kejiang <kj.ma@intel.com>
* fix lint issues
Signed-off-by: Ma, Kejiang <kj.ma@intel.com>
* fix issues for format, type checking
Signed-off-by: Ma, Kejiang <kj.ma@intel.com>
* remove test code
Signed-off-by: Ma, Kejiang <kj.ma@intel.com>
* fix issues after rebase
Signed-off-by: Ma, Kejiang <kj.ma@intel.com>
---------
Signed-off-by: Ma, Kejiang <kj.ma@intel.com>
* feat: add LongCat provider support
- Add LongCat to SystemProviderIds enum
- Add LongCat provider logo and configuration
- Configure API endpoints and URLs based on official docs
- Add two models: LongCat-Flash-Chat and LongCat-Flash-Thinking
- Update provider mappings for proper integration
The LongCat provider uses OpenAI-compatible API format and supports
up to 8K tokens output with daily free quota of 500K tokens.
Signed-off-by: LeaderOnePro <leaderonepro@outlook.com>
* feat: add migration for LongCat provider
- Add migration version 158 for LongCat provider
- Ensure existing users get LongCat provider on app update
- Follow standard migration pattern for simple provider additions
Signed-off-by: LeaderOnePro <leaderonepro@outlook.com>
---------
Signed-off-by: LeaderOnePro <leaderonepro@outlook.com>
- Introduced a new utility function to determine if a tool is an agent tool, simplifying the tool selection logic in MessageTool.
- Refactored MessageAgentTools to improve rendering logic and added an UnknownToolRenderer for better handling of unrecognized tools.
- Updated BashOutputTool to remove unnecessary Card components, enhancing layout consistency.
- Improved overall code clarity and maintainability by reducing redundancy and adhering to existing patterns.
* feat(toolUsePlugin): separate provider-defined tools from prompt tools in context
- Enhanced the `createPromptToolUsePlugin` function to distinguish between provider-defined tools and other tools, ensuring only non-provider-defined tools are saved in the context.
- Updated the handling of tools in the transformed parameters to retain provider-defined tools while removing others.
- Improved error handling in `ToolExecutor` by logging tool and tool use details for better debugging.
- Refactored various components to use `NormalToolResponse` instead of `MCPToolResponse`, aligning with the new response structure across multiple message components.
* refactor(toolUsePlugin): streamline tool handling in createPromptToolUsePlugin
- Updated the `createPromptToolUsePlugin` function to improve type handling for tools, ensuring proper type inference and reducing the use of type assertions.
- Enhanced clarity in the separation of provider-defined tools and prompt tools, maintaining functionality while improving code readability.
* refactor(ToolExecutor): remove debug logging for tool and tool use
- Removed console logging for tool and tool use details in the ToolExecutor class to clean up the code and improve performance. This change enhances the clarity of the code without affecting functionality.
Add Breadcrumbs and HorizontalScrollContainer components to enhance agent selection UI. Remove redundant agent name display since it's now shown in the breadcrumb chip. Improve layout with better overflow handling and responsive design.
* refactor(tools): enhance descriptions for knowledge and web search tools
- Updated the descriptions for the knowledgeSearchTool and webSearchTool to provide clearer context on their functionality.
- Improved the formatting of prepared queries and relevant links in the descriptions to enhance user understanding.
- Added information on how to use the tools with additional context for refined searches.
* fix:format lint
- Add max-width to agent name tag in ChatNavbar
- Adjust header padding in AgentSettingsPopup
- Replace span with Ellipsis component for agent names to handle overflow
This PR correctly addresses an event propagation issue where clicking the close button on an error alert was unintentionally triggering the parent click handler (which opens the detail modal).
* refactor(notes): improve notes management with local state and file handling
- Replace UUID-based IDs with SHA1 hash of file paths for better consistency
- Remove database storage for notes tree, use local state management instead
- Add localStorage persistence for starred and expanded states
- Improve cross-platform path normalization (replace backslashes with forward slashes)
- Refactor tree operations to use optimized in-memory operations
- Enhance file watcher integration for better sync performance
- Simplify notes service with direct file system operations
- Remove database dependencies from notes tree management
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
* Revert "Merge remote-tracking branch 'origin/main' into refactor/note"
This reverts commit 389386ace8, reversing
changes made to 4428f511b0.
* fix: format error
* refactor: noteservice
* refactor(notes): 完成笔记状态从localStorage向Redux的迁移
- 将starred和expanded路径状态从localStorage迁移到Redux store
- 添加版本159迁移逻辑，自动从localStorage迁移现有数据到Redux
- 优化NotesPage组件，使用Redux状态管理替代本地localStorage操作
- 改进SaveToKnowledgePopup的错误处理和验证逻辑
- 删除NotesTreeService中已废弃的localStorage写入函数
- 增强组件性能，使用ref避免不必要的依赖更新
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
* fix: ci
* feat(notes): add in-place renaming for notes in HeaderNavbar
- Implemented an input field for renaming the current note directly in the HeaderNavbar.
- Added handlers for title change, blur, and key events to manage renaming logic.
- Updated the breadcrumb display to accommodate the new title input.
- Enhanced styling for the title input to ensure seamless integration with the existing UI.
This feature improves user experience by allowing quick edits without navigating away from the notes list.
* Update NotesEditor.tsx
---------
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: kangfenmao <kangfenmao@qq.com>
- Extract repeated div styling into reusable InfoTag component
- Add agent name to the info items display
- Replace inline styles with tailwind classes for consistency
- Add 'selected' key to common section in all language files (en-us, zh-cn, zh-tw, el-gr, es-es, fr-fr, ja-jp, pt-pt, ru-ru)
- Fix CLAUDE.md documentation to use correct 'yarn sync:i18n' command
- Resolve '[to be translated]' placeholders with proper localized translations
- Ensure consistency across all supported languages
Fixes missing i18n key error: [I18N] Missing key: common.selected
Reset the topic fulfilled state when switching between sessions to ensure proper state management. Also remove redundant state update from SessionItem's onPress handler.
Remove the redundant dispatch of setTopicFulfilled in messageThunk since it's now handled in SessionItem. Add visual indicators for pending and fulfilled states in SessionItem to improve user feedback.
- Add getErrorMessage utility for consistent error message formatting
- Enhance addAgent to return Result type for better error handling
- Add disallowEmptySelection to form dropdowns
- Reset loading state on errors in AgentModal
- Rename getAgentAvatar to getAgentDefaultAvatar for clarity
- Add EmojiAvatarWithPicker component for emoji selection
- Update AgentLabel to support both default and emoji avatars
- Add AvatarSetting component for avatar configuration
- Modify agent configuration schema to support emoji avatars
Add modelFilter parameter to SelectApiModelPopup to exclude embedding, rerank and text-to-image models from selection. This ensures only appropriate models are shown based on agent type requirements.
- Refactor useUpdateAgent to return both updateAgent and updateModel functions
- Update all components using useUpdateAgent to use the new hook structure
- Improve model selection by reusing SelectAgentModelButton component
- Add pagination support to useApiModels hook
Add a confirmation step before deleting a session, including a tooltip showing the keyboard shortcut. Uses a timer to automatically cancel the confirmation after 3 seconds.
- Add support for anthropicApiHost configuration in providers
- Improve model filtering for Anthropic-compatible providers
- Add isAnthropicModel function to validate Anthropic models
- Update ClaudeCode service to support compatible providers
- Enhance logging and error handling in API routes
- Fix model transformation and validation logic
- Add `anthropicApiHost` field to Provider type - Update provider config
and migration to set Anthropic endpoints - Add UI for configuring
Anthropic API Host in provider settings - Update SDK client logic to use
Anthropic API Host when available - Add i18n strings for Anthropic API
Host configuration
Move the lucide icon color rule into the media query block for consistency.
Add AlertTriangleIcon to caution text in AgentModal for better visual warning.
* feat: improve content protection during file operations
- Add validation for knowledge base configuration before saving
- Enhance error handling for note content reading
- Implement content backup and restoration during file rename
- Add content verification after rename operations
- Improve user feedback with specific error messages
* fix: format check
---------
Co-authored-by: 自由的世界人 <3196812536@qq.com>
fix(websearch): handle blocked domains conditionally in web search configurations
- Updated the handling of blocked domains in both Google Vertex and Anthropic web search configurations to only include them if they are present, improving robustness and preventing unnecessary parameters from being passed.
- Commented out all references to the 'cherryin' provider in configuration files.
- Updated the version in the persisted reducer from 157 to 158.
- Added migration logic to remove 'cherryin' from the state during version 158 migration.
* Fix slash menu Shift+Enter newline
* fix: enable Shift+Enter newline in rich editor with slash commands
Fixed an issue where users couldn't create new lines using Shift+Enter when
slash command menu (/foo) was active. The problem was caused by globa
keyboard event handlers intercepting all Enter key variants.
Changes:
- Allow Shift+Enter to pass through QuickPanel event handling
- Add Shift+Enter detection in CommandListPopover to return false
- Implement fallback Shift+Enter handling in command suggestion render
- Remove unused import in AppUpdater.ts
- Convert Chinese comments to English in QuickPanel
- Add test coverage for command suggestion functionality
---------
Co-authored-by: Zhaokun Zhang <zhaokunzhang@Zhaokuns-Air.lan>
* feat: improve content protection during file operations
- Add validation for knowledge base configuration before saving
- Enhance error handling for note content reading
- Implement content backup and restoration during file rename
- Add content verification after rename operations
- Improve user feedback with specific error messages
* fix: format check
---------
Co-authored-by: 自由的世界人 <3196812536@qq.com>
- Updated references from 'agents' to 'assistants' across various components and hooks.
- Changed i18n keys to reflect the new terminology for better clarity.
- Removed the deprecated agents slice and integrated its functionality into the assistants slice.
- Adjusted UI components to align with the new naming conventions for assistant presets.
fix(websearch): handle blocked domains conditionally in web search configurations
- Updated the handling of blocked domains in both Google Vertex and Anthropic web search configurations to only include them if they are present, improving robustness and preventing unnecessary parameters from being passed.
- Commented out all references to the 'cherryin' provider in configuration files.
- Updated the version in the persisted reducer from 157 to 158.
- Added migration logic to remove 'cherryin' from the state during version 158 migration.
* Fix slash menu Shift+Enter newline
* fix: enable Shift+Enter newline in rich editor with slash commands
Fixed an issue where users couldn't create new lines using Shift+Enter when
slash command menu (/foo) was active. The problem was caused by globa
keyboard event handlers intercepting all Enter key variants.
Changes:
- Allow Shift+Enter to pass through QuickPanel event handling
- Add Shift+Enter detection in CommandListPopover to return false
- Implement fallback Shift+Enter handling in command suggestion render
- Remove unused import in AppUpdater.ts
- Convert Chinese comments to English in QuickPanel
- Add test coverage for command suggestion functionality
---------
Co-authored-by: Zhaokun Zhang <zhaokunzhang@Zhaokuns-Air.lan>
- Change hover background color and add shadow in AgentItem
- Use cn utility for className in SessionItem
- Update height and background color for active state in SessionItem
- Introduce SelectAgentModelButton component for agent model selection
- Add SelectApiModelPopup for displaying and selecting API models
- Implement apiModelAdapter to convert API models to adapted format
- Add model filtering by agent type in agentSession utils
- Update model select components to use new API model selection
Update component and type names from 'advance' to 'advanced' for consistency and correct spelling. This includes renaming the file and all related references in the codebase.
- Updated logging statements across various modules to provide more structured and detailed information.
- Changed log levels from info to debug for less critical messages to reduce log clutter.
- Enhanced error logging to include relevant context such as agentId, sessionId, and model details.
- Standardized log messages to follow a consistent format, improving readability and maintainability.
Ensure proper handling of undefined values for agent name and description by making state types optional. Also update the updateName function to handle optional name input.
- Updated CodeToolsPage to include checks for supported endpoint types for various CLI tools.
- Added 'cherryin' to GEMINI_SUPPORTED_PROVIDERS and updated CLAUDE_SUPPORTED_PROVIDERS to include it.
- Improved logic for determining model compatibility with selected CLI tools, enhancing overall functionality.
- Introduced a helper function to escape strings for AppleScript to ensure proper command execution.
- Updated terminal command definitions to utilize the new escape function, improving compatibility with special characters.
- Adjusted command parameters to use double quotes for directory paths, enhancing consistency and reliability.
* refactor(reasoning): simplify reasoning time tracking by removing unused variables and logic
- Removed hasStartedThinking and reasoningBlockId variables as they are no longer needed.
- Updated onThinkingComplete callback to eliminate final_thinking_millsec parameter, streamlining the function.
* refactor(thinking): streamline thinking millisecond tracking and update event handling
- Removed unused thinking_millsec parameter from onThinkingComplete and adjusted related logic.
- Updated AiSdkToChunkAdapter to simplify reasoning-end event handling by removing unnecessary properties.
- Modified integration tests to reflect changes in thinking event structure.
* fix: 添加 seedThink 标签以支持新的模型识别
* Enable reasoning for SEED-OSS models
- Add SEED-OSS model ID check to reasoning exclusion logic
- Include SEED-OSS models in reasoning model detection
* fix: 更新 reasoning-end 事件处理以使用最终推理内容
Use agent's model information instead of session's to maintain consistency across the application. The model ID format changed from "sessionId:modelId" to "provider:modelId" and the model object is now constructed using the actual model details from the agent.
- Added isNewApiProvider function to streamline checks for 'new-api' and 'cherryin' providers.
- Updated ApiClientFactory, providerConfig, and various components to utilize isNewApiProvider for improved readability and maintainability.
- Refactored conditional checks across multiple files to replace direct string comparisons with the new utility function.
- Adjusted padding in TabsBar and Navbar components to enhance spacing.
- Updated ItemRenderer and Sortable components to accept itemStyle prop for custom styling.
- Changed NotesSidebar scroll behavior from 'smooth' to 'instant'.
- Modified MCP server card width to 100% for better responsiveness.
- Set wrapperStyle and itemStyle to 100% width in McpServersList for consistent item display.
* feat(minapps): support temporary minapps
* feat(settings): use openSmartMinApp with app logo to open docs sites
* refactor(icons): replace styled img with tailwind
* feat(tab): tighten types
* feat(tab): use minapps cache and log missing entries
* test(icons): update MinAppIcon snapshot to reflect class and attrs
* fix(assistant): enforce id requirement when updating assistant
Ensure assistant id is always provided when updating assistant properties by making it a required field in the update payload. This prevents potential bugs where updates might be applied to wrong assistants.
* refactor(useAssistant): simplify updateAssistant callback by removing redundant id
Update InputbarTools to use simplified callback signature
- Implement double-click to edit session names directly in the list
- Add loading state during save operation
- Update useInPlaceEdit hook to support async operations and saving state
- Adjust styling to accommodate new edit input field
- Adjusted the width of the CardContainer to dynamically calculate based on viewport width.
- Changed the layout of the McpServersList from grid to list, with a vertical orientation and updated styling for list items.
- Introduced new HTML files for the privacy policy in English and Chinese.
- Implemented a PrivacyPopup component to display the privacy policy within the application.
- The popup dynamically loads the appropriate language based on user settings and includes options to accept or decline the policy.
- Introduced a new IPC channel for quitting the application.
- Updated ipc.ts to handle the App_Quit channel, allowing the app to quit when invoked.
- Added corresponding quit method in the preload API for client-side access.
- Fixed a minor URL check in WindowService to ensure proper navigation handling.
- Added detailed JSDoc comments for clarity on tool input types, including ReadToolInput, TaskToolInput, BashToolInput, and others.
- Introduced new input types such as ListMcpResourcesToolInput and ReadMcpResourceToolInput to expand functionality.
- Improved existing types to ensure better documentation and usability for developers.
- Introduced new tools: EditTool, MultiEditTool, BashOutputTool, NotebookEditTool, and ExitPlanModeTool.
- Updated MessageTool to support new tool types.
- Enhanced ReadTool to handle output as an array of text outputs.
- Improved type definitions to accommodate new tools and their inputs/outputs.
- Add text color classes to modal title and list items for better visibility
- Apply background and border styling to modal content
- Use modal hook pattern for consistency
- Translated and reorganized Russian language JSON for tooling and permissions.
- Removed deprecated MCP and tool settings components.
- Introduced new AgentToolingSettings component to manage tooling permissions and MCP servers.
- Updated AgentSettings index to reflect new tooling settings structure.
- Enhanced agent configuration schema to include permission modes with default values.
The antd Select component was replaced with a custom Select component from @heroui/react to improve consistency with the design system. This change also simplifies the model selection logic by removing the need for manual option mapping.
- Move tool selection from essential settings to dedicated "Pre-approved tools" tab
- Update terminology from "Allowed tools" to "Pre-approved tools" for clarity
- Add new AgentToolSettings component with enhanced card-based layout
- Include warning alert about pre-approved tools bypassing review
- Update all language files with new terminology and translation keys
- Add i18n sync guidance to CLAUDE.md development commands
- Add MCP server configuration UI for agent settings
- Update agent and session forms to include MCP server selection
- Fix MCP API service logging and tools handling
- Add Chinese localization for MCP settings
- Update type definitions to support MCP server arrays
This enables agents to use MCP (Model Control Protocol) servers
as additional tools and capabilities in their execution context.
- Add filter support to useApiModel hook for provider-specific models
- Improve ApiModelLabel with customizable classNames for styling
- Update ChatNavbar to use filtered models for agents
Add comprehensive tool management UI allowing users to select which tools are pre-approved for agents and sessions. Includes multi-select dropdowns with tool descriptions, proper validation, and internationalization support across 10+ languages.
- Add tool selection UI to AgentModal, SessionModal, and AgentEssentialSettings
- Extend BaseAgentForm and related types with allowed_tools field
- Implement tool validation and filtering logic
- Add i18n support for tool selection labels and descriptions
- Include visual chip-based display for selected tools
- Introduced WebFetchTool for fetching web content with specified prompts and URLs.
- Updated MessageTool to include WebFetch in the tool rendering options.
- Enhanced BashTool and TaskTool to improve display of input and output information.
- Refactored GenericTools for better parameter display and consistency across tools.
- Adjusted types to include WebFetchTool input and output definitions.
- Update DexieMessageDataSource to delete files when count reaches zero and deleteIfZero is true
- Add deleteIfZero parameter to MessageDataSource interface and all implementations
- Modify updateFileCountV2 thunk to pass deleteIfZero parameter through DbService
- Replace manual message loading logic with loadTopicMessagesThunk for better caching
- Remove unused imports and local state management
- Simplify useEffect dependencies and loading flow
* fix(assistant): update translate assistant content handling for QwenMT model
- Adjusted content assignment in getDefaultTranslateAssistant to use store settings only when the model is not a QwenMT model, ensuring correct translation behavior.
* lint err
* refactor(assistant): encapsulate content handling logic for translation
- Introduced a new function, getTranslateContent, to streamline content assignment in getDefaultTranslateAssistant.
- This change improves readability and maintains correct translation behavior for QwenMT models.
* format code
- Add Redux selector to check for existing messages in store
- Always reload messages to Redux when session data is available
- Add effect to restore messages when component mounts if missing from Redux
- Replace manual Redux logic with `useTopicMessages` hook for consistent message loading behavior
- Add `deleteMessages` method to message data sources with proper block and file cleanup
- Update `DbService` to delegate batch deletion to appropriate data source implementations
- Prevent unnecessary message reloads by checking existing messages before loading session messages
- Implement LRU cache and throttled persistence for streaming agent messages to reduce backend load
- Add streaming state detection and proper cleanup for complete messages to improve performance
- Delete persistExchange method from all data sources and DbService
- Remove unused Topic import and MessageExchange type dependencies
- Simplify agent session existence check to validate sessionId directly
- Make getRawTopic required in MessageDataSource interface
- Add comprehensive solution documentation for status persistence and streaming state
- Implement message update functionality in AgentMessageDataSource for agent sessions
- Remove redundant persistAgentExchange logic to eliminate duplicate saves
- Streamline message persistence flow to use appendMessage and updateMessageAndBlocks consistently
- Modify AgentMessageDataSource.appendMessage to save messages to backend immediately instead of waiting for response completion
- Add proper error handling and logging for message persistence operations
- Create comprehensive test documentation covering V2 database service scenarios
- Integrate V2 implementations for message operations (save, update, delete, clear) with feature flag control
- Add topic creation fallback in DexieMessageDataSource when loading non-existent topics
- Create integration status documentation tracking completed and pending V2 migrations
- Update Topic type to include TopicType enum for proper topic classification
* refactor(Tools): replace MCPToolResponse with NormalToolResponse in message tools and add new agent tools
- Updated MessageKnowledgeSearch, MessageMemorySearch, and MessageWebSearch components to use NormalToolResponse.
- Refactored MessageTool to handle NormalToolResponse and simplified tool rendering logic.
- Introduced new agent tools: BashTool, GlobTool, GrepTool, ReadTool, SearchTool, TaskTool, and TodoWriteTool with corresponding types and renderers.
- Enhanced type safety by updating tool response types in the codebase.
* fix(i18n): Auto update translations for PR #10303
* chunk type
* refactor(migration): renumber migration steps after removing step 155
Remove unused migration step 155 and renumber subsequent steps to maintain sequence
* fix(store): prevent mutation of assistant presets by using spread operator
* fix(store): ignore ts-2589 false positives and refactor preset updates
Refactor assistant preset updates to use forEach instead of map for consistency
Add ts-ignore comments for TypeScript false positives in store operations
* Fix tool result handling and session creation flow
- Populate toolName in tool-result chunks from contentBlockState
- Add onSessionCreated callback to SessionModal for post-creation actions
- Return created session from useSessions hook and update SWR cache optimistically
* Fix toolName reference in ClaudeCode message transformation
- Correctly reference toolName from contentBlockState using blockKey instead of block.tool_use_id
- Ensure proper tool result chunk generation when handling assistant messages
- Maintain consistent data structure for tool call processing
* Fix toolName reference and add stream event logging
- Correct toolName lookup to use tool_use_id instead of blockKey in tool-result chunks
- Add debug logging for stream event handling
- Update contentBlockState key to use event.content_block.id for tool_use events
* Add debug logging for message content blocks
- Log each content block when processing user or assistant messages
- Maintain existing switch case logic for text block handling
- Improve debugging visibility for multi-block message processing
* get toolName
* chore: bump version to 1.7.0-alpha.1
* fix(getSdkClient): add authToken to anthropic client initialization for claude code
* Update transform.ts
* Refactor logging levels in transform.ts and adjust JSON body parser configuration in app.ts
* refactor(sessions): simplify session creation by removing modal and using direct button
The SessionModal component was removed and replaced with a direct button click handler that creates a session using the agent data. Also added error handling to display an alert when session fetching fails.
* feat(agents): add api server check and warning for agent features
- Add api server enabled check in multiple components
- Show warning alert when api server is disabled
- Add dismissable warning in AgentSection
- Disable send button when api server is disabled
- Add iknow state to store dismissed warnings
* feat(i18n): add warning message for enabling API server
Add warning message in multiple languages to inform users they need to enable API server to use agent features
* feat(sessions): make session creation async and set active session
Dispatch active session id after successful creation to ensure UI reflects current state
* feat(sessions): add session waiting state and improve deletion handling
- Add sessionWaiting state to track updating/deleting sessions
- Extract updateSession logic into separate hook
- Improve session deletion with waiting state and fallback session selection
- Disable session items during deletion to prevent duplicate actions
* feat(i18n): add error message for last session deletion
Add error message to prevent deletion of the last session in all supported languages
* fix(i18n): Auto update translations for PR #10096
* fix(i18n): Auto update translations for PR #10096
* feat(tools): add WriteTool and update tool rendering logic
- Introduced WriteTool for handling file writing operations.
- Updated MessageAgentTools to include the new WriteTool in the tool renderers.
- Refactored existing tools to streamline rendering and improve code clarity.
- Enhanced BashTool and TaskTool to better display input and output information.
---------
Co-authored-by: GitHub Action <action@github.com>
Co-authored-by: suyao <sy20010504@gmail.com>
Co-authored-by: icarus <eurfelux@gmail.com>
Co-authored-by: Vaayne <liu.vaayne@gmail.com>
- Add sessionWaiting state to track updating/deleting sessions
- Extract updateSession logic into separate hook
- Improve session deletion with waiting state and fallback session selection
- Disable session items during deletion to prevent duplicate actions
- Add api server enabled check in multiple components
- Show warning alert when api server is disabled
- Add dismissable warning in AgentSection
- Disable send button when api server is disabled
- Add iknow state to store dismissed warnings
The SessionModal component was removed and replaced with a direct button click handler that creates a session using the agent data. Also added error handling to display an alert when session fetching fails.
- Log each content block when processing user or assistant messages
- Maintain existing switch case logic for text block handling
- Improve debugging visibility for multi-block message processing
- Correct toolName lookup to use tool_use_id instead of blockKey in tool-result chunks
- Add debug logging for stream event handling
- Update contentBlockState key to use event.content_block.id for tool_use events
- Correctly reference toolName from contentBlockState using blockKey instead of block.tool_use_id
- Ensure proper tool result chunk generation when handling assistant messages
- Maintain consistent data structure for tool call processing
- Populate toolName in tool-result chunks from contentBlockState
- Add onSessionCreated callback to SessionModal for post-creation actions
- Return created session from useSessions hook and update SWR cache optimistically
Refactor assistant preset updates to use forEach instead of map for consistency
Add ts-ignore comments for TypeScript false positives in store operations
- Add UI for managing accessible paths in agent settings
- Improve error handling and loading states in agent components
- Update type definitions for better type safety
- Remove outdated comments and fix styling issues
Simplify agent type label handling by replacing the map with a direct switch statement. Also update related type references and array type assertions for consistency.
- Introduced a new method in ConfigManager to generate and retrieve a unique client ID.
- Updated AppUpdater to include the client ID in the request headers alongside the user agent.
Clean up agents store by removing deprecated AgentEntity-related code that is no longer used. This simplifies the store structure as we're moving away from redux.
* fix: support leadingspace to avoid normal text
* Close QuickPanel when no search results found
Add automatic closing of QuickPanel when search yields no results for
single-select input triggers, preventing users from getting stuck with
empty result lists.
* fix: reopen quick panel while editing trigger text
* fix: hide quick trigger hints when disabled
* Update zh-tw.json
- Move 'label' from agent type to root type section
- Add new prompt settings and type-related translations
- Update Chinese translations for consistency
Extract ModelLabel component into standalone ApiModelLabel and rename useModels hook to useApiModels for better clarity. Update all references to use the new names. This improves code organization and maintainability.
- Make model prop optional in ModelAvatar component
- Ensure models array always returns an array in useModels hook
- Export SelectorProps type for reuse
- Add getProviderNameById utility function
- Introduce ModelLabel component for displaying model info
- Update AgentEssentialSettings to include model selection dropdown
- Replace SettingsInline with more flexible SettingsItem component
- Add SettingsContainer for consistent layout
- Remove redundant styled components in favor of shared components
- Move common components (AgentLabel, SettingsTitle, SettingsInline) to shared file
- Update CSS to use @layer base for better organization
- Fix agent type label translation key in AgentModal
- Add agent type label utility function
- Introduce new AgentPromptSettings component for managing agent prompts
- Move prompt-related functionality from AgentEssentialSettings to new component
- Add avatar display to essential settings
- Improve layout structure and styling for both settings components
- Rename built_in_tools field to tools for consistency
- Add type field to Tool schema (builtin/mcp/custom)
- Consolidate tool handling in BaseService with listMcpTools method
- Remove unused CreateSessionResponse and related schemas
- Clean up unused imports and dead code in session handlers
- Unify agent and session tool resolution logic
- Add shared Anthropic utilities package with OAuth and API key client creation
- Implement provider-specific message routing alongside existing v1 API
- Enhance authentication middleware with priority handling (API key > Bearer token)
- Add comprehensive auth middleware test suite with timing attack protection
- Update session handling and message transformation for Claude Code integration
- Improve error handling and validation across message processing pipeline
- Standardize import formatting and code structure across affected modules
This establishes the foundation for Claude Code OAuth authentication while maintaining
backward compatibility with existing API key authentication methods.
Move update agent functionality from useAgent and useAgents hooks into a dedicated useUpdateAgent hook to improve code organization and maintainability
Add translations for accessible paths section in session settings across multiple languages
Move accessible_paths object to consistent location in JSON structure
Add placeholder for essential settings in all language files
Move agent editing functionality from inline modal to a dedicated settings popup component for better maintainability and separation of concerns. The new implementation provides a more structured settings interface with essential agent configuration options.
- Move Assistants and Agents components to dedicated folders
- Split TopicsTab into separate Topics and SessionsTab components
- Add activeTopicOrSession state handling in runtime store
- Update tab switching logic to support both topics and sessions
- Clean up and optimize component imports and exports
- Split AssistantsTab into separate components (Assistants and Agents)
- Add SectionName component for better UI organization
- Remove unused tabs ('agents' and 'sessions') from chat type
- Clean up imports and type definitions
- Automatically select first agent if none active after loading
- Automatically select first session per agent if none active after loading
- Prevent empty selection states in UI components
- Extract actual model ID from session model string and resolve model details
- Use resolved model name, provider, and group when available instead of defaults
- Remove redundant API model handling in getModelName function
- Use split and pop to isolate the model name after the colon
- Fall back to the full model string if no colon is present
- Maintain provider and group identifiers for model object consistency
- Replace complex fallback chain for provider display name with direct provider name access
- Add console.log for model debugging in getModelName function
- Add ApiModel type import and update function signature to accept ApiModel
- Return formatted name using provider_name field for API models
- Maintain backward compatibility for legacy models by looking up provider in store
- Replace static "No active agent" message with a spinner and loading text
- Integrate react-i18next for translation of loading message
- Adjust animation timing and styling for smoother loading state transition
- Replace static conditional rendering with Framer Motion animations for no-agent and session states
- Animate session list items with staggered entrance and exit transitions
- Add loading spinner animation with fade effect
- Apply motion to session creation button with delayed entrance
- Wrap `SessionMessages` and `SessionInputBar` in `useMemo` to prevent unnecessary re-renders
- Refactor `AgentSessionMessages` to use shared layout components and message grouping
- Extract common styled components to `shared.tsx` for reuse across message components
- Add provider_name field to ApiModel schema and transformation logic
- Update model options to include providerName for better display
- Improve provider label fallback chain in model transformation
- Fix agent hook to use proper SWR key and conditional fetching
- Enhance option rendering with better truncation and provider display
Move accessible paths functionality from session modal to agent modal,
add validation requiring at least one path, and update form handling to
inherit agent paths in sessions.
Replaced EventEmitter-based agent streaming with ReadableStream for
better compatibility with AI SDK patterns. Modified
SessionMessageService to return stream/completion pair instead of event
emitter, updated HTTP handlers to use stream pumping, and added IPC
contract for renderer-side message persistence.
- Remove redundant agentId checks as they're handled by the API client
- Add consistent error formatting using formatErrorMessageWithPrefix
- Update error messages for all session operations
- Extract common option components to shared.tsx for reuse
- Make useModels filter parameter optional
- Update SessionModal to use real model data from API
Consolidate error formatting functions (formatAgentServerError and formatAxiosError) into error.ts utility file to improve code organization and maintainability
- Remove unused persistence tracking variables in message handler
- Simplify finalizeResponse logic by removing unnecessary checks
- Change 'finish' event type to 'complete' for consistency
- Add debug logging for streaming events
- Clean up dead code and improve readability
- Extract getServersFromRedux to shared utility getMCPServersFromRedux
- Implement 5-minute TTL cache for MCP servers and providers
- Reduce redundant Redux store queries in API server
- Improve response times for frequently accessed data
Change unsupported provider error to return undefined
Replace thrown error with empty object return and update function
signature to allow undefined return type for unsupported providers
- Change condition from text.trim() === '' to text.length === 0
- Users can now delete whitespace characters (spaces, tabs, newlines) without accidentally deleting attached files
- File deletion only occurs when text is completely empty
- Maintains existing functionality for file deletion when text is truly empty
Fixes: Blank character input causes backspace to incorrectly delete attached images
Signed-off-by: zhaokun <zhaokun_zhang@icloud.com>
- Change default message id from -1 to 77777 in useSession
- Remove schema validation for session response temporarily
- Add proper content parsing for agent session messages
Implement optimistic UI updates when creating new messages to improve perceived performance. The changes include cloning the current session data, adding a draft message immediately, and handling rollback on error.
Implement agent session messages display component and track active topic/session state
Add AgentSessionMessages component and integrate with chat view
Update topic and session selection to set active state in store
Add new state field and action to track whether the user is viewing topics or sessions in the chat interface. This enables proper UI state management when switching between views.
The previous implementation unnecessarily spread the previous state when only the result is needed. This simplifies the mutation logic while maintaining the same behavior.
- Implement SessionModal component for creating/editing sessions
- Replace Button wrapper with fragment in SessionItem for cleaner styling
- Add translation support and proper form handling for session creation
Remove redundant sessionId parameter from updateSession methods since it's already included in the UpdateSessionForm. This makes the API more consistent and reduces potential for mismatched IDs.
Restructure the component organization by moving AgentModal.tsx into a dedicated agent subdirectory under Popups. This improves maintainability by grouping related agent components together.
Update all import paths to reflect the new location. The component functionality remains unchanged.
- Rename Container to ButtonContainer for consistency
- Add activeSessionId state to track active sessions per agent
- Implement Sessions and SessionItem components with loading state
- Add session selection and deletion functionality
Move sessions rendering logic to a separate component and modify useSessions hook to work with agentId instead of full agent entity. This improves code organization and simplifies the hook interface.
