Graphiti MCP Server

🚀 Turing MCP Server 🧠

A powerful knowledge graph server designed for AI agents, integrating Neo4j and the Model Context Protocol (MCP).

🚀 Quick Start

This is a powerful knowledge graph server designed for AI agents, integrating Neo4j and the Model Context Protocol (MCP).

✨ Features

• 🔄 Dynamic knowledge graph management with Neo4j.
• 🤖 Seamless integration with OpenAI models.
• 🔌 Support for MCP (Model Context Protocol).
• 🐳 Ready for Docker deployment.
• 🎯 Custom entity extraction functionality.
• 🔍 Advanced semantic search.

📦 Installation

Prerequisites

• Docker and Docker Compose.
• Python 3.10 or higher.
• OpenAI API key.

Quick Start 🚀

1. Clone the repository:

git clone https://github.com/gifflet/graphiti-mcp-server.git
cd graphiti-mcp-server

2. Set up environment variables:

cp .env.sample .env

3. Edit the .env file as needed:

# Required for LLM operations
OPENAI_API_KEY=your_openai_api_key_here
MODEL_NAME=gpt-4o

4. Start the service:

docker compose up

🔧 Technical Details

Neo4j Configuration 🗄️

The default configuration for Neo4j is as follows:

• Username: neo4j
• Password: demodemo
• URI: bolt://neo4j:7687 (within the Docker network)
• Memory settings are optimized for the development environment.

Docker Environment Variables 🐳

You can run with environment variables directly:

OPENAI_API_KEY=your_key MODEL_NAME=gpt-4o docker compose up

💻 Usage Examples

Cursor IDE Integration 🖥️

1. Configure Cursor to connect to Graphiti:

{
  "mcpServers": {
    "Graphiti": {
      "url": "http://localhost:8000/sse"
    }
  }
}

2. Add Graphiti rules to Cursor's user rules (see graphiti_cursor_rules.md).
3. Start an agent session in Cursor.

🏗️ Architecture

The server mainly consists of two parts:

• A graph storage using the Neo4j database.
• The Turing MCP Server, responsible for API and LLM operations.

🤝 Contribution

Contributions are welcome! Please feel free to submit a Pull Request.

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• The Neo4j team for the excellent graph database.
• OpenAI for the powerful LLM models.
• The MCP community for the protocol specification.