ladams9999/memory-mcp-server

Memory MCP Server

An MCP (Model Context Protocol) server that provides long-term memory capabilities for AI agents, enabling persistent storage, retrieval, and semantic search of contextual memories across sessions.

I've created this project because I've found myself jumping between multiple models, agents, and even apps in the course of working on my projects. For the coding agents, you can nominally maintain a file to retain memory, but I've been using tools that don't have that access, and also working with data I don't want saved publicly as part of the work. This is part of the reason I've started with local first architecture (Ollama + ChromaDB).

Features

• Store memories with contextual information
• Retrieve memories by context
• Semantic search across stored memories
• ChromaDB backend for persistent storage
• Ollama integration for local embeddings

Setup

1. Clone this repository
2. Install UV package manager (installation guide)
3. Copy .env.example to .env and configure environment variables
4. Install dependencies: uv sync. This should also create the virtual environment
5. If Ollama is not running, start the service: ollama serve
6. Pull the embedding model: ollama pull mxbai-embed-large
7. Run the server: uv run mcp-memory-server

Uvicorn uses websockets in a way that has been deprecated in websockets >= 14.0. You may see a warning to this effect when you start the server.

Configuration

See .env.example for configuration options.

Status

🚧 This project is currently under development 🚧

This is an MVP implementation focusing on ChromaDB + Ollama integration.

Development

After setting up the project and activating the virtual environment, use the UV package manager to manage dependencies, run the server, and execute tests:

# Install all dependencies including dev dependencies
uv sync --dev

# Run the server
uv run mcp-memory-server

# Run all tests
uv run pytest

# Run specific test file
uv run pytest tests/unit/test_memory_service.py

# Run tests with coverage
uv run pytest --cov=src/mcp_memory_server --cov-report=html --cov-report=term-missing

Agent Integration

To use the MCP Memory Server as a memory backend for your agent, you need to configure your agent or development environment to connect to the server's MCP endpoint.

VS Code Settings

Create or update your mcp.json file (location: VS Code user settings or agent config directory):

{
  "servers": {
    "MCP Memory Server": {
      "url": "http://localhost:8139/mcp/",
      "type": "http"
    }
  },
  "inputs": []
}

You may run into a situation where VS Code limits the number of tools you make available to its agents, and will start chucking errors when you try to use them if you are over this limit. You can resolve this by going to the tool list and unchecking the tools you aren't using until you get under that limit.

Warp Settings

{
  "MCP Memory Server": {
    "url": "http://localhost:8139/mcp/"
  }
}

• Ensure the MCP server is running locally (uv run mcp-memory-server).
• Update the endpoint if running on a different host or port.

Notes:

• Agents may require a restart after changing environment variables or config files.
• If running in a container or remote environment, update the endpoint to match your network setup.

The MCP server must be running and accessible from your agent's environment.

• If you use Docker or remote servers, update the endpoint accordingly.
• For more advanced agent integrations, refer to your agent's documentation for MCP memory configuration options.