Neura Market
MindGeniusAI logo

MindGeniusAI

Free

Auto generate MindMap with ChatGPT ![GitHub Repo stars](https://img.shields.io/github/stars/xianjianlf2/MindGeniusAI?style=social)

FreeFree tier
Type
Open Source
Visit Website GitHub (40 stars)

About MindGeniusAI

MindGeniusAI is a self-hostable, open-source AI agent that converts documents (particularly PDFs) into editable mind maps. Built for privacy-conscious users, it runs entirely on your own infrastructure via a single Docker image, ensuring your files and API keys never leave your machine. Unlike typical cloud-based mind map tools, MindGeniusAI features a real multi-step AI agent (not just a prompt wrapper) that plans, retrieves information, draws mind maps live, and can surgically edit existing maps by node ID. It supports multiple AI providers including OpenAI, Claude, DeepSeek, and Kimi, and includes built-in RAG (retrieval-augmented generation) for PDF chunking and embedding. The tool provides a two-way canvas where users can manually edit nodes and the AI builds on those changes. All tool calls are visible in real-time.

Key Features

Self-hosted via one Docker image; files and API keys never leave your machine
Real multi-step AI agent with visible tool calls (plan, retrieve, draw, edit)
Surgical edits: modify specific nodes by ID without full re-render
Two-way canvas: manual node edits are recognized and built upon by the AI
Reads PDFs: chunk → embed → in-memory RAG retrieval (~100 lines, no LangChain)
Multi-provider support: OpenAI, Claude, DeepSeek, Kimi, any compatible endpoint
Bring-your-own-API-key stored only in your browser
Source-available under BSL 1.1 license, hackable codebase

Pros & Cons

Pros
  • Complete privacy: runs on your own machine, no data sent to third parties
  • Real AI agent with visible reasoning steps, not a one-shot prompt
  • Supports multiple popular LLM providers (OpenAI, Claude, DeepSeek, Kimi)
  • Surgical editing preserves manual tweaks to mind maps
  • Two-way collaboration between user and AI on the same canvas
  • Open source (BSL 1.1) and hackable with simple RAG implementation
  • Free to use with your own API keys
Cons
  • Requires technical setup: Docker and API key configuration
  • Performance depends on your own hardware and chosen model provider
  • Live demo first load may take ~30 seconds to wake
  • Primarily designed for PDFs; other document formats may not be supported as natively
  • No managed cloud version: user must self-host or run locally

Best For

Privacy-conscious knowledge workers who cannot upload sensitive documents to cloud servicesSelf-hosters and developers wanting full control over their mind mapping pipelineTeams with sensitive documents (legal, research, internal specs) requiring local processingCreating and iteratively editing mind maps from PDF reports or research papersEducational scenarios where students want to visualize document content with custom AI models

FAQ

How is MindGeniusAI different from typical cloud mind-map tools?
Unlike cloud SaaS that uploads your documents to their servers and holds your model API key in their backend, MindGeniusAI runs entirely on your own infrastructure via Docker. Your files stay local and your API key lives only in your browser.
What AI models can I use with MindGeniusAI?
It supports OpenAI, Claude, DeepSeek, Kimi, and any compatible API endpoint. You bring your own API key.
Can I run MindGeniusAI locally without internet?
You need internet access to call the AI provider's API, but the application and your documents run locally. The source code is self-hostable with Docker.
Does MindGeniusAI store my documents or data?
No. Your PDFs are chunked and embedded in memory only, nothing is written to disk on the server, and there is no telemetry on your content. Your API key is stored only in your browser.
Can I edit the mind map manually while the AI works?
Yes, the two-way canvas allows you to rename, add, drag, or delete nodes by hand. The AI sees your changes and builds on them like a collaborator, rather than regenerating the entire map.