next-ai-draw-io version history

- Fixed window close behavior on macOS/Windows for draw.io editing. - Added a dropdown for all draw.io themes in the settings panel. - Ensured access code config updates correctly when settings dialog opens. - Improved AI Model Configuration sidebar layout for better usability.

## What's New in v0.4.15 ### New Features - **Personal My Templates Library**: Added personal My Templates library alongside Quick Examples for saving and reusing diagram templates (#773) - **XML+SVG Export**: Added xmlsvg export option for diagrams (#761) - **RPM Package for Linux**: Added rpm target to electron-builder for Linux distribution (#806) ### Improvements - **Tool Demo Updates**: Updated tool demos on the English about page (#598) - **MCP Tool Description**: Improved create_new_diagram tool description to prevent misuse (#803) - **Chat Lobby Panels**: Show all enabled panels on chat lobby instead of only templates (#802) ### Bug Fixes - **QvQ Vision Model Support**: Allow QvQ (Qwen Visual QA) models to use image input (#808) - **Qwen3.5 Vision Model Support**: Allow Qwen3.5 models to use image input (#800) - **CPU Usage Optimization**: Reduce CPU usage during large XML streaming (#801) - **Undo/Redo Restoration**: Remove redundant status call after autosave to restore undo/redo (#780) - **MCP Security - Localhost Binding**: Bind MCP server HTTP to 127.0.0.1 only (#787) - **MCP Security - Body Size Limit**: Add 10MB body size limit to MCP HTTP endpoints (#791) - **

## What's New in v0.4.14 ### New Features - **Novita AI Provider**: Added Novita AI as a new LLM provider with OpenAI-compatible API and open-source model support (#758) ### Improvements - **GLM Vision Model Check**: Added checks for GLM text and visual model naming conventions (#741) - **MiniMax M2.7 Model**: Upgraded MiniMax default model to M2.7 with high-speed variant (#737) ### Bug Fixes - **MCP Server CORS Security**: Restricted CORS to same-origin only, preventing external sites from accessing the MCP server (#757) - **Settings Dialog Overflow**: Fixed settings dialog overflowing outside the viewport with thin scrollbar styling (#743) - **Electron Startup on Windows/Linux**: Resolved Electron startup failures caused by port conflicts on Windows and Linux (#709) - **IndexedDB Recovery**: Fixed IDB closing recovery and restored diagram-only sessions after refresh (#648) ### Developer Experience - **CI Cleanup**: Removed `npm audit` step from CI that was blocking unrelated PRs (#744) ### Contributors - @Alex-wuhu - @Biki-dev - @DayuanJiang - @octo-patch - @sbilly ### Downloads | Platform | File | Notes | |----------|------|-------| | macOS (Intel) | [Next-AI-Draw.io

## What's New in v0.4.13 ### New Features - **Chinese AI Provider Support**: Added support for MiniMax, GLM (Zhipu), Qwen (Alibaba), Kimi (Moonshot), and Qiniu AI providers (#722) - **Custom System Message**: Added custom system message setting for AI personalization (#728) - **API Key Load Balancing**: Added API key load balancing for providers (#676) - **Material Design Icons**: Added Material Design Icons shape library (#688) - **MCP PNG/SVG Export**: Added PNG/SVG export to MCP server (#687) - **Stop Button**: Added stop button to cancel AI generation (#660) ### Improvements - **Ollama Cloud Support**: Added Ollama Cloud support with Base URL and API Key configuration (#692) - **Client-provided Ollama Base URL**: Supports client-provided base URL for Ollama (#664) - **Qwen 3.5 Plus Vision Model**: Added Qwen 3.5-plus vision model support (#706) - **Kimi K2.5 Image Support**: Enabled image support for Kimi K2.5 model (#670) - **Traditional Chinese Locale**: Added Traditional Chinese (zh-Hant) locale support (#650) - **Language-aware Responses**: Added language-aware response rules with English fallback (#641) - **API Keys Discoverability**: Added "API Keys & Models" link in

## What's New in v0.4.12 ### New Features - **Offline Draw.io Support in Desktop App**: The Electron desktop app now bundles draw.io (v29.3.5) for offline support - no internet connection required to use the diagram editor (#629) - **VLM-based Diagram Validation**: Added visual validation for AI-generated diagrams using Vision Language Models with auto-retry (up to 3 attempts) when validation fails (#602) ### Bug Fixes - **MCP Auto-redirect**: Fixed confusing "No session" infinite spinner when accessing MCP server URL without session parameter - now auto-redirects to active session (#634) - **Quota Bypass for Bedrock Users**: Fixed issue where users with their own Bedrock credentials were still subject to quota limits (#621) - **Provider Selection Border**: Fixed incomplete border rendering when a provider is selected in the model configuration dialog (#617) - **Security Update**: Upgraded wrangler to ^4.60.0 to address CVE-2026-0933 (command injection vulnerability) (#628) ### Improvements - **MCP UI Modernization**: Updated MCP server UI with project logo, modern dialog styles with backdrop blur, animations, and DM Sans font (#636) - **Base URL Hint**: Added URL format hint

## What's New in v0.4.11 ### New Features - **Server-side Multi-provider/Model Support**: Administrators can configure multiple AI models server-side that are available to all users without requiring personal API keys. Supports environment variable or config file setup with a new `/api/server-models` endpoint (#583) - **Google Cloud Vertex AI**: Added Vertex AI as a new provider using GCP service account authentication - no API key needed (#574) - **Enter/Ctrl+Enter Send Setting**: Choose between Enter or Ctrl+Enter to send messages in chat (#550) - **Electron Menu i18n**: Electron desktop menu now automatically displays in English, Chinese, or Japanese based on system language (#596) ### Bug Fixes - **Edit Diagram JSON Parsing**: Fixed inconsistent quote escaping in edit_diagram JSON where LLM generates patterns like `y="-20\"` causing JSON parsing failures (#601) - **Private URL Support for Reverse Proxy**: Added `ALLOW_PRIVATE_URLS` env var (defaults to `true`) to allow localhost and private IPs for self-hosted reverse proxy setups (#600) - **MCP Edit After Create**: Fixed error "You must call get_diagram first" when calling `edit_diagram` immediately after `create_new_diagr

## What's New in v0.4.10 ### New Features - **URL Content Extraction**: Extract and analyze content from URLs directly in chat (#514) - **ModelScope Support**: Added ModelScope as a new AI provider (#521) - **Proxy Settings**: Configure HTTP/HTTPS proxy for API requests in Desktop app (#537) - **Windows Code Signing**: Added SignPath code signing for Windows builds (#531) ### Bug Fixes - **URL Extraction Loading State**: Fix loading state stuck on failure (#524) - **Doubao Models**: Use OpenAI provider for Doubao multimodal models (#519) - **Subdirectory Deployment**: Add NEXT_PUBLIC_BASE_PATH build arg (#518) - **Missing Dependency**: Add missing nanoid dependency (#528) - **SignPath Config**: Specify artifact-configuration-slug for SignPath (#533) ### Improvements - **Testing Infrastructure**: Add Vitest and Playwright testing framework (#512) - **MCP Server**: Fix author and repository metadata (#529) - **About Pages**: Remove usage limits from about pages (#520) ### Contributors - @Biki-dev - @broBinChen - @DayuanJiang - @yrk111222 - @zhoujie0531 ### Downloads | Platform | File | Notes | |----------|------|-------| | macOS (Intel) | [Next-AI-Draw.io-0.4.10.dmg](https:

## What's New in v0.4.9 ### New Features - **Chat Session History**: Added chat session history with IndexedDB persistence (#500) ### Bug Fixes - **Save Panel Fix**: Remove draw.io native save button to prevent duplicate dialogs (#507) - **Chat Button**: Disable new chat button during streaming to prevent errors (#501) - **MCP Spinner**: Prevent stuck spinner by initializing blank session state (#494) - **Toast Position**: Move toast notifications to bottom-left for better UX (#489) - **Docker**: Fix invalid YAML syntax in docker-compose.yml (#498) ### Improvements - **Code Refactor**: Extract ToolCallCard and ChatLobby components for cleaner code (#502) - **Save Feedback**: Show success toast after saving diagram (#484) ### Documentation - **Cline MCP**: Add Cline MCP configuration instructions (#488) ### Contributors - @DayuanJiang - @RohitChavan16 - @yupeng0512 ### Downloads | Platform | File | |----------|------| | macOS (Intel) | [Next-AI-Draw.io-0.4.9.dmg](https://github.com/DayuanJiang/next-ai-draw-io/releases/download/v0.4.9/Next-AI-Draw.io-0.4.9.dmg) | | macOS (Apple Silicon) | [Next-AI-Draw.io-0.4.9-arm64.dmg](https://github.com/DayuanJiang/next-ai-draw-io/rel

## What's New in v0.4.8 🎉 **Happy New Year!** This release focuses on bug fixes and quality-of-life improvements. Stay tuned for the next version which will bring the **session history feature**! --- ### New Features - **EdgeOne Pages Provider**: Add EdgeOne Pages as AI provider option (#456) - **Copy Tool Calls**: Add copy button for tool call blocks (#463) - **Unvalidated Models Toggle**: Add toggle to show unvalidated models in model selector (#413) - **MCP Private Deployments**: Add DRAWIO_BASE_URL env for private deployments (#467) ### Bug Fixes - **Loading Flash Fixes**: Hide Draw.io loading flash with placeholder, prevent flash of example panel (#481, #482) - **Faster Page Restore**: Faster message restore and skip panel animation on refresh (#483) - **DeepSeek Image Error**: Improve image not supported error detection for DeepSeek (#474) - **MCP Graceful Shutdown**: Add graceful shutdown to prevent zombie processes (#477) - **Cascade Delete**: Implement cascade delete for children and edges in edit_diagram (#451) - **Locale Redirect**: Restore locale redirection using Next.js middleware (#462) - **Dark Mode Logo**: Switch logo file based on dark mode (#425) - **Resp

## What's New in v0.4.7 ### New Features - **i18n Model Config**: Add translations for model configuration UI (#368) - **Dev XML Simulator**: Add streaming simulator for UI debugging (#385) ### Bug Fixes - **Electron Settings Persistence**: Fix settings not persisting after app restart by using fixed port (#415) - **OpenAI Reasoning Display**: Fix reasoning/thinking blocks not showing for OpenAI models (#370) - **OpenAI-Compatible Proxies**: Use Chat Completions API for better proxy compatibility (#382) - **Model Compatibility**: Rename edit_diagram type field to operation for better model compatibility (#402) - **HTML Escape in XML**: Fix parse errors by escaping HTML in XML attribute values (#386) - **Progressive Streaming**: Enable diagram rendering during streaming (#380) - **Continuation Retry**: Add retry limit for truncated diagrams (#372) - **UI Fixes**: Reset test button on provider switch, scrollable model list (#411, #395, #396) ### Improvements - **Config Panel Redesign**: Refined minimal aesthetic for configuration panels (#384) - **AI SDK v6**: Upgrade to AI SDK v6.0.1 with simplified Langfuse integration (#369, #375) - **Code Refactoring**: Extract diagram tool