MindGraph is an open-source, API-first prototype for building AI-powered, graph-based CRM solutions. It targets developers and researchers looking for a flexible, extensible framework to manage and query knowledge graphs through natural language, offering a foundation for custom data solutions.

How It Works

MindGraph utilizes a Flask backend with an in-memory graph data structure for entities and relationships, managed via a schema.json file. It supports dynamic integrations, allowing AI models to process natural language inputs into structured graph data. The system exposes RESTful API endpoints for CRUD operations, search, and triggering custom integrations, with an optional frontend for visualization.

Quick Start & Requirements

• Install: git clone https://github.com/yoheinakajima/MindGraph.git, cd MindGraph, poetry install
• Prerequisites: Python 3.6+, Poetry. Requires an OpenAI API key set in a .env file.
• Run: poetry run python main.py
• Docs: https://github.com/yoheinakajima/MindGraph

Highlighted Details

• API-first design with RESTful endpoints for entity and relationship management.
• Dynamic integration system for extending functionality with AI and external systems.
• Schema-driven knowledge graph creation using schema.json for AI input processing.
• Supports multiple database backends including in-memory, NexusDB, NebulaGraph, and FalkorDB.

Maintenance & Community

The project is a proof-of-concept and the author explicitly states they do not maintain it, inviting others to take over. Contact via X/Twitter: https://twitter.com/yoheinakajima.

Licensing & Compatibility

MIT License. Permissive for commercial use and closed-source linking.

Limitations & Caveats

The project is a proof-of-concept with no active maintenance. The author is seeking a maintainer. The in-memory database is not recommended for production. The frontend search form is noted as a quick showcase and not ideal for natural language queries.