Next AI Draw.io is a Next.js web application that combines artificial intelligence with draw.io diagram creation, enabling users to generate, modify, and enhance diagrams through natural language commands. The application is written in TypeScript and hosted at next-ai-drawio.jiang.jp, serving as a productivity tool that bridges the gap between conversational AI and visual diagram design.
The core functionality centers on LLM-powered diagram creation, allowing users to describe what they want in natural language and have the AI generate corresponding draw.io XML diagrams. The application supports image-based diagram replication, where users can upload existing diagrams or images for the AI to replicate and enhance automatically. Additionally, users can upload PDF documents and text files to extract content and generate diagrams from those sources. For supported models including OpenAI o1 and o3, Gemini, and Claude, the application displays the AI's reasoning process, providing transparency into how diagrams are being constructed.
The platform includes comprehensive diagram history and version control, allowing users to track all changes and restore previous versions before AI editing. An interactive chat interface enables real-time refinement of diagrams through conversation with the AI. The application offers specialized support for cloud architecture diagrams across AWS, GCP, and Azure, and includes the ability to create animated connectors between diagram elements for enhanced visualization.
The most active labels tracked are bug with 140 items, enhancement with 103 items, and help wanted with 41 items.
The application supports multiple AI providers including ByteDance Doubao, AWS Bedrock, OpenAI, Anthropic, Google AI, Google Vertex AI, Azure OpenAI, Ollama, OpenRouter, AIHubMix, DeepSeek, SiliconFlow, ModelScope, SGLang, and Vercel AI Gateway. Administrators can configure multiple server-side models through environment variables or JSON configuration files, and an admin panel accessible via password allows management of server settings, models, access codes, features, and quota without manual environment file editing.
The application can be deployed across multiple platforms including Tencent EdgeOne Pages, Vercel, and Cloudflare Workers. A desktop application is available for Windows, macOS, and Linux from the releases page. The project also includes MCP Server integration, allowing the application to work with AI agents like Claude Desktop, Cursor, and VS Code through the Model Context Protocol, enabling users to create diagrams directly through these interfaces.
The underlying technology stack uses Next.js for the frontend framework, Vercel AI SDK for streaming responses and multi-provider support, and react-drawio for diagram representation and manipulation. The application stores diagrams as XML that can be rendered in draw.io, with the AI processing user commands to generate or modify this XML accordingly.
