Releases: coderchatwang/CoderChat
1.2.2
Release v1.2.2
This release brings comprehensive improvements to the theme system, enhanced chat features, better Windows platform support, and several important fixes.
🎉 New Features
Theme & UI Improvements
- Unified theme management – Added
VoidThemeDataProviderto centrally manage theme state. - Theme variable support – Replaced hardcoded black with
currentColorfor full theme switching compatibility. - VoidSwitch component optimization – Replaced hardcoded colors with VSCode theme variables.
- New color variable – Added
inputBackgroundRevertfor reverse background color scenarios.
Chat Enhancements
- Message timestamps – Chat messages now display their creation time.
- Automatic session projectId update – Session projectId is now kept up to date automatically.
- Paste state synchronization – Fixed an issue where the send button remained disabled after pasting; paste state is now properly synced.
Recent Items Management
- Delete support – Added the ability to delete individual items from the recent files/projects list.
Network Proxy Support
- Proxy configuration – Added proxy support for Anthropic and Google SDKs.
- HTTP configuration options – Upgraded Anthropic SDK and added HTTP configuration capabilities.
- System message handling fix – Fixed system message handling in anthropic-style.
Windows Platform Support
- Close confirmation dialog – Enabled the close confirmation dialog by default on Windows to prevent accidental closure.
🐛 Bug Fixes
- Fixed paste state synchronization, resolving the issue where the send button became disabled.
- Fixed delayed pid handling in
node-ptyand improvedSharedArrayBufferhandling. - Fixed i18n formatting issues.
- Fixed anthropic-style system message handling.
📦 Dependency Updates
- Upgraded
@vscode/windows-process-treeto^0.7.0 - Upgraded type definition dependencies
- Upgraded Anthropic SDK
- Added
floating-uiimport
📚 Documentation Updates
- Added download links to README
- Updated developer information and added ARM64 build script documentation
- Updated version number to 1.2.2
🔧 Build & Development
- Windows build task – Added Void build task to support Windows and cross-platform builds.
- ARM64 build script – Added
build-release-arm64.ps1for ARM64 release builds. - Installation script – Added
install.shfor Node.js environment setup.
Thank you to all contributors and users for your support! Welcome to download the latest version and try it out.
Note: This is a preview version featuring the latest functionalities. Open source is not easy — to avoid the situation faced by the original Void project and ensure the sustainability of the current project, the source code will be released later, depending on how much attention the project receives. A good project relies on continuous feedback and improvement. If you like this project, please help promote it to those in need, and when you encounter issues, let me know via GitHub Issues. If you would like to join the ongoing development of this project, you can contact the author by email to obtain the latest unreleased source code. Staying true to the original spirit of "coding for love," the author will continue to release the latest functional versions under Releases for developers who need them.
1.2.1
Version 1.2.1 contains breaking changes. Before using version 1.2.1, you need to clear the configuration data in the CoderChat directory under your user directory. For Mac users, delete /Users/*/Library/Application Support/CoderChat; for Windows users, delete C:\Users*\AppData\Roaming\CoderChat.
Highlights
- Fixed multi-window conversation loss issue.
- Enhanced API calling compatibility: Supports native Anthropic SDK calls (Gemini still uses OpenAI-compatible format).
- Centralized theme management: Unified state manager makes theme switching more stable and fixes the freeze issue when entering light theme for the first time.
- Configurable AI response language: You can now specify the language AI uses for replies; added Markdown cache toggle for better rendering performance.
- Improved new user onboarding: Language selection added to the setup flow for a friendlier first-time experience.
- Support for AGENTS.md.
New Features
API & Model Configuration
- Upgraded Anthropic SDK with support for custom HTTP configuration.
- Added support for calling Anthropic models using the native Anthropic SDK format (Gemini still uses OpenAI format).
- Added a "Compatible API Provider" setting option and labels for OpenAI-compatible configuration items in the UI.
Theme & UI
- Added
VoidThemeDataProviderto centrally manage theme state, improving theme switching experience. - Provider in settings panel now supports expand/collapse.
- Upgraded dependency libraries to align with the latest version of VS Code.
Multi-window & Sync
- Supports real-time session data synchronization across multiple windows.
Internationalization (i18n)
- Added i18n support for the edit model dialog and the model dialog.
- Added language selection UI to the onboarding page, supporting multiple languages.
- Supports filtering conversation threads by project; enhanced language selection during onboarding.
AI Response Settings
- Added AI response language setting to specify the language used by AI.
- Added Markdown cache toggle to improve rendering performance.
Gather Mode
- Added Gather mode confirmation dialog with i18n support.
- Enhanced Gather mode prompts for better code collection.
Configuration Files
- Supports AGENTS.md.
Bug Fixes
- Fixed delayed process ID issue in
node-pty, improvedSharedArrayBufferhandling. - Fixed application ID error on Windows ARM64.
- Fixed key collision in dropdown menus and disabled searching symlinks by default.
- Fixed errors when using the ask tool.
- Fixed TypeScript compilation configuration and some type definitions.
- Fixed the type definition of
IApplicationStorageMainService(createDecoratorgeneric parameter).
Refactors
- Refactored the "Add Model" functionality in settings panel, split into two sections for better layout.
- Removed the
specialToolFormatproperty from default model configuration to simplify config structure.
Style Adjustments
- Adjusted various UI spacing to improve visual experience.
Note: This is a preview version featuring the latest functionalities. Open source is not easy — to avoid the situation faced by the original Void project and ensure the sustainability of the current project, the source code will be released later, depending on how much attention the project receives. A good project relies on continuous feedback and improvement. If you like this project, please help promote it to those in need, and when you encounter issues, let me know via GitHub Issues. If you would like to join the ongoing development of this project, you can contact the author by email to obtain the latest unreleased source code. Staying true to the original spirit of "coding for love," the author will continue to release the latest functional versions under Releases for developers who need them.
1.2.0
Feature Update Changelog
New Features
- Added Gather mode confirmation UI with internationalization support
- Renamed
.voidrulestoAGENTS.mdfor AI context configuration - Added label display for OpenAI Compatible series configuration items
- Added provider expand/collapse functionality
- Added multi‑window session data synchronization
- Added internationalization support for the model dialog
- Added support for calling Anthropic SDK using Anthropic format (Note: Gemini requires OpenAI format)
Improvements & Optimizations
- Removed the
specialToolFormatattribute from the default model configuration - Refactored settings page: split the model addition functionality into top and bottom sections
- Adjusted UI spacing styles
- Added
compatibleApiProviderNamesimport to the VoidOnboarding component - Enhanced Gather mode prompts
- Replaced "Void" characters (brand‑related)
- Updated version number and configuration information
Bug Fixes
- Fixed dropdown menu key collision issue
- Disabled the default value for searching symbolic links
Note: This is a preview version featuring the latest functionalities. Open source is not easy — to avoid the situation faced by the original Void project and ensure the sustainability of the current project, the source code will be released later, depending on how much attention the project receives. A good project relies on continuous feedback and improvement. If you like this project, please help promote it to those in need, and when you encounter issues, let me know via GitHub Issues. If you would like to join the ongoing development of this project, you can contact the author by email to obtain the latest unreleased source code. Staying true to the original spirit of "coding for love," the author will continue to release the latest functional versions under Releases for developers who need them.
1.1.0
New Features
State Management Optimization
- Zustand State Management Integration
- Introduced Zustand for fine-grained state management
- Improved state update performance and responsiveness
History Message Pagination
- History Message Pagination Loading
- Support for paginated display of history messages
- Load 30 more messages at a time
- Optimized performance for large amounts of history messages
Project-level Thread Management
-
Project-based Thread Filtering
- Added project-based thread filtering in sidebar
- Support viewing and managing conversation threads by project
-
Workspace Change Listener
- Added workspace change listener
- Automatically update projectId for empty threads
Message Display Control
- Reset Visible Messages Setting
- Added
resetVisibleOnSendsetting - Option to reset visible messages when sending
- Provided default value configuration
- Added
Performance Optimization
- Markdown Parsing Cache
- Added Markdown parsing cache mechanism
- Significantly improved message rendering performance
Tool Method Enhancement
- Array Field Description Support
- Tool methods support defining array field descriptions
- Note: Currently only definition is supported, not yet applied
Vision Support
- Image Formatting Support
- Added vision support check in LLM messages
- Improved formatting for image messages
HTML Processing
- stripHtml Support
- Added HTML tag stripping functionality
- Used for cleaning HTML tags in message content
Bug Fixes
Scroll Position Preservation
- Fixed scroll position when loading more messages
- Maintain current scroll position when loading more history messages
- Improved user experience
Other Fixes
- General fixes
Other Updates
Branding Update
- Void character replacement
- Brand-related character replacement updates
Note: This is a preview version featuring the latest functionalities. Open source is not easy — to avoid the situation faced by the original Void project and ensure the sustainability of the current project, the source code will be released later, depending on how much attention the project receives. A good project relies on continuous feedback and improvement. If you like this project, please help promote it to those in need, and when you encounter issues, let me know via GitHub Issues. If you would like to join the ongoing development of this project, you can contact the author by email to obtain the latest unreleased source code. Staying true to the original spirit of "coding for love," the author will continue to release the latest functional versions under Releases for developers who need them.
1.0.0
Changelog - Updates Since 2026/04/14
Chat Functionality
- Added ask_user_question tool
- Added xml_escape tool
- Support for sending images with tool notifications
- Added OpenAI-style tool format configuration
Internationalization (i18n)
- Added internationalization framework for Void Chat
- Added Chinese default values (voidChatICn18n)
- Tool names in chat interface support Chinese localization
- Merged i18n configs into unified useVoidChatI18n hook
Developer Mode and Debugging
- Added developer mode section in settings UI
- Added showJsonDebug toggle to control JSON debug area visibility
- Display role and model information
- Improved debug title and JSON content display
UI/UX Improvements
- UI optimization for deep thinking section
- Edit prompt functionality
- Display retry errors in UI
- Optimized retry logic
- OpenAI Compatible shows source URL
Vision Support
- Added supportsVision support
Build and Configuration
- Added Node.js memory limit to build scripts to fix OOM issues
- Added Mac startup script
Branding
- Renamed from Void to CoderChat
- Updated application icons with transparent backgrounds