|
Hey,
I built an open-source tool that lets you build draw.io diagrams using LLMs. Unlike image generators, this manipulates the actual draw.io XML. This enables a hybrid workflow: prompt the AI to build a base architecture, tweak it manually on the canvas, then ask the AI to refine it again. Key features:
- Cloud Ready: Native support for AWS, GCP, and Azure icon sets (not generic boxes).
- Dynamic: Supports animated connectors and vector sketching (e.g., "draw a cat").
- Model Agnostic: Works with OpenAI, Anthropic, Bedrock, and Ollama. I'd love feedback on the app, or ideas for v2 (e.g., add MCP feature). Any ideas are appreciated. Thanks! Demo: https://next-ai-drawio.jiang.jp/
Repo: https://github.com/DayuanJiang/next-ai-draw-io |