A-MEM MCP Server

A Memory Control Protocol (MCP) server for the Agentic Memory (A-MEM) system - a flexible, dynamic memory system for LLM agents.

Overview

The A-MEM MCP Server provides a RESTful API wrapper around the core Agentic Memory (A-MEM) system, enabling easy integration with any LLM agent framework. The server exposes endpoints for memory creation, retrieval, updating, deletion, and search operations.

A-MEM is a novel agentic memory system for LLM agents that can dynamically organize memories without predetermined operations, drawing inspiration from the Zettelkasten method of knowledge management.

Key Features

• 🔄 RESTful API for memory operations
• 🧠 Dynamic memory organization based on Zettelkasten principles
• 🔍 Intelligent indexing and linking of memories
• 📝 Comprehensive note generation with structured attributes
• 🌐 Interconnected knowledge networks
• 🧬 Continuous memory evolution and refinement
• 🤖 Agent-driven decision making for adaptive memory management

Installation

1. Clone the repository:

git clone https://github.com/Titan-co/amem-mcp-server.git
cd amem-mcp-server

2. Install dependencies:

pip install -r requirements.txt

3. Start the server:

uvicorn server:app --host 0.0.0.0 --port 8000 --reload

API Endpoints

Create Memory

• Endpoint: POST /memories
• Description: Create a new memory note
• Request Body:

  {
    "content": "string",
    "tags": ["string"],
    "category": "string",
    "timestamp": "string"
  }
  

Get Memory

• Endpoint: GET /memories/{id}
• Description: Retrieve a memory by ID

Update Memory

• Endpoint: PUT /memories/{id}
• Description: Update an existing memory
• Request Body:

  {
    "content": "string",
    "tags": ["string"],
    "category": "string",
    "context": "string",
    "keywords": ["string"]
  }
  

Delete Memory

• Endpoint: DELETE /memories/{id}
• Description: Delete a memory by ID

Search Memories

• Endpoint: GET /memories/search?query={query}&k={k}
• Description: Search for memories based on query
• Query Parameters:
  • query: Search query string
  • k: Number of results to return (default: 5)

Configuration

The server can be configured through environment variables:

• OPENAI_API_KEY: API key for OpenAI services
• LLM_BACKEND: LLM backend to use (openai or ollama, default: openai)
• LLM_MODEL: LLM model to use (default: gpt-4)
• EMBEDDING_MODEL: Embedding model for semantic search (default: all-MiniLM-L6-v2)
• EVO_THRESHOLD: Number of memories before triggering evolution (default: 3)

Documentation

Interactive API documentation is available at:

• Swagger UI: http://localhost:8000/docs
• ReDoc: http://localhost:8000/redoc

References

Based on research paper: A-MEM: Agentic Memory for LLM Agents