Memgpt MCP Server

🚀 MemGPT MCP Server

This is a TypeScript-based MCP server that implements a memory system for large language models (LLMs) and provides the ability to chat with different LLM providers. The server maintains conversation history and supports the use of multiple different LLM providers.

🚀 Quick Start

The MemGPT MCP Server is a powerful tool for interacting with various LLM providers. It offers a range of features to manage conversations and memory effectively.

✨ Features

Tools

• chat - Send a message to the current LLM provider

  • Accepts a single message parameter
  • Supports multiple providers (OpenAI, Anthropic, OpenRouter, Ollama)

• get_memory - Retrieve conversation history

  • Optional limit parameter to specify the number of memories to retrieve
  • Use limit: null to retrieve all memories without a limit
  • Returns memories in chronological order with timestamps

• clear_memory - Clear the conversation history

  • Deletes all stored memories

• use_provider - Switch between different LLM providers

  • Supports OpenAI, Anthropic, OpenRouter, and Ollama
  • Persists the provider selection

• use_model - Switch the model for the current provider

  • Supports provider-specific models:
    • Anthropic Claude model series:
      • Claude 3 series:
        • claude-3-haiku: Fastest response, suitable for tasks like customer support and content moderation
        • claude-3-sonnet: Balanced performance, suitable for general use
        • claude-3-opus: Advanced model, suitable for complex reasoning and high-performance tasks
      • Claude 3.5 series:
        • claude-3.5-haiku: Faster response and more cost-effective
        • claude-3.5-sonnet: Better performance with computer interaction capabilities
    • OpenAI: Supported models include 'gpt-4o', 'gpt-4o-mini', and 'gpt-4-turbo'
    • OpenRouter: Supports any model in the 'provider/model' format (e.g., 'openai/gpt-4', 'anthropic/claude-2')
    • Ollama: Supports any locally available model (e.g., 'llama2', 'codellama')
  • Persists the model selection

📦 Installation

Development

Install dependencies:

npm install

Build the server:

npm run build

Development with automatic rebuild:

npm run watch

Configuration for Claude Desktop

To use on Claude Desktop, add the server configuration to:

• MacOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%/Claude/claude_desktop_config.json

Example configuration:

{
  "servers": {
    "your-server-name": {
      "type": "memgpt",
      "host": "localhost",
      "port": 12345,
      "apiKey": "your-api-key"
    }
  }
}

🔧 Technical Details

Environment Variables

Before running the server, you need to set the following environment variables:

• OPENAI_API_KEY - OpenAI API key
• ANTHROPIC_API_KEY - Anthropic API key
• OLLAMA_SERVER_URL (optional) - URL of the Ollama server, defaults to http://localhost:11434

Debugging

The MCP Inspector tool can be used for debugging as it communicates via standard input/output. This is useful for monitoring and adjusting the server behavior.

📚 Documentation

Recent Updates

• Added support for Claude 3 and 3.5 series models
• Optimized the flexibility and usability of the memory retrieval function