An AI-powered desktop mind-mapping app.
Turn ideas, documents, and conversations into living mind maps.
Describe what you're thinking in the floating chat panel and watch a mind map grow on the canvas. The assistant streams its progress live, and can keep refining the map — adding branches, restructuring nodes, and answering questions about what it built.
Drop in what you already have and get a structured map back:
| Format | Source |
|---|---|
| research papers, reports, ebooks | |
| DOCX / PPTX / XLSX | Office documents |
| Markdown | notes and drafts |
| Web URL | articles and pages |
| Plain text | anything on your clipboard |
A fast, free-form canvas built on React Flow. Drag nodes anywhere, restyle them in the style panel, auto-layout when things get messy, and drive everything from the keyboard.
Bring your own API key — switch providers without changing how you work.
| Provider | Models | Extras |
|---|---|---|
| 通义千问 (百炼) | qwen-turbo / qwen-plus / qwen-max / qwen-long | vision + image generation |
| Kimi Code | Kimi K2.5 / Kimi K2 | — |
| MiniMax | MiniMax M2.7 / M2.5 / M2.1 / M2 | image generation |
| DeepSeek (V4) | DeepSeek V4 Flash / V4 Pro | — |
| OpenCode Go | GLM-5.x · Kimi K2.7 Code · Kimi K2.6 … | — |
Extend what the assistant can reach through the Model Context Protocol — a built-in server catalog with brand icons, extensible via settings.json:
npm install
npm run dev # start the app (Vite + Electron)
npm run build # type-check, bundle, and package with electron-builderThen open Settings, pick a provider, and paste in your API key.
github.com/FourWindff/MindLane · fourwindff.github.io/MindLane