Memgpt

🚀 MemGPT MCP Server

A TypeScript-based MCP server implementing an LLM memory system. It provides tools for chatting with different LLM providers while maintaining conversation history.

🚀 Quick Start

The MemGPT MCP Server is a TypeScript-based application that offers an LLM memory system. It enables users to chat with various LLM providers and keeps track of the conversation history.

✨ Features

Tools

• chat - Send a message to the current LLM provider

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

• get_memory - Retrieve the conversation history

  • Optional limit parameter to specify the number of memories to retrieve
  • Use limit: null to retrieve 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
  • Keeps the provider selection persistent

• use_model - Switch to a different model of the current provider

  • Supports provider-specific models:
    • Anthropic Claude models:
      • Claude 3 series:
        • claude-3-haiku: Fastest response time, suitable for tasks like customer support and content moderation
        • claude-3-sonnet: Balanced performance, suitable for general purposes
        • claude-3-opus: Advanced model, suitable for complex reasoning and high-performance tasks
      • Claude 3.5 series:
        • claude-3.5-haiku: Enhanced speed and cost-effectiveness
        • claude-3.5-sonnet: Superior performance with computer interaction capabilities
    • OpenAI: 'gpt-4o', 'gpt-4o-mini', 'gpt-4-turbo'
    • OpenRouter: Any provider/model format (e.g., 'openai/gpt-4', 'anthropic/claude-2')
    • Ollama: Any locally available model (e.g., 'llama2', 'codellama')
  • Keeps the model selection persistent

📦 Installation

With Claude Desktop

When using with Claude Desktop, add the server configuration:

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

{
  "mcpServers": {
    "letta-memgpt": {
      "command": "/path/to/memgpt-server/build/index.js",
      "env": {
        "OPENAI_API_KEY": "your-openai-key",
        "ANTHROPIC_API_KEY": "your-anthropic-key",
        "OPENROUTER_API_KEY": "your-openrouter-key"
      }
    }
  }
}

Environment Variables

• OPENAI_API_KEY - Your OpenAI API key
• ANTHROPIC_API_KEY - Your Anthropic API key
• OPENROUTER_API_KEY - Your OpenRouter API key

Debugging

Since the MCP server communicates via standard input/output, debugging can be challenging. We recommend using MCP Inspector:

npm run inspector

The Inspector will open a new window to assist you with debugging.

🔧 Technical Details

Development

Install dependencies:

npm install

Build the server:

npm run build

Development with auto-rebuild:

npm run watch

📚 Documentation

Recent Updates

• Enhanced support for Claude 3 series models
• Improved infinite memory retrieval functionality