MCP Memory Server: Scalable Long-Term AI Memory

MCP Memory Server

This server implements long-term memory capabilities for AI assistants using mem0 principles, powered by PostgreSQL with pgvector for efficient vector similarity search.

Features

• PostgreSQL with pgvector for vector similarity search
• Automatic embedding generation using BERT
• RESTful API for memory operations
• Semantic search capabilities
• Support for different types of memories (learnings, experiences, etc.)
• Tag-based memory retrieval
• Confidence scoring for memories
• Server-Sent Events (SSE) for real-time updates
• Cursor MCP protocol compatible

Prerequisites

1. PostgreSQL 14+ with pgvector extension installed:

# In your PostgreSQL instance:
CREATE EXTENSION vector;

2. Node.js 16+

Setup

1. Install dependencies:

npm install

2. Configure environment variables: Copy .env.sample to .env and adjust the values:

cp .env.sample .env

Example .env configurations:

# With username/password
DATABASE_URL="postgresql://username:password@localhost:5432/mcp_memory"
PORT=3333

# Local development with peer authentication
DATABASE_URL="postgresql:///mcp_memory"
PORT=3333

3. Initialize the database:

npm run prisma:migrate

4. Start the server:

npm start

For development with auto-reload:

npm run dev

Using with Cursor

Adding the MCP Server in Cursor

To add the memory server to Cursor, you need to modify your MCP configuration file located at ~/.cursor/mcp.json. Add the following configuration to the mcpServers object:

{
  "mcpServers": {
    "memory": {
      "command": "node",
      "args": [
        "/path/to/your/memory/src/server.js"
      ]
    }
  }
}

Replace /path/to/your/memory with the actual path to your memory server installation.

For example, if you cloned the repository to /Users/username/workspace/memory, your configuration would look like:

{
  "mcpServers": {
    "memory": {
      "command": "node",
      "args": [
        "/Users/username/workspace/memory/src/server.js"
      ]
    }
  }
}

The server will be automatically started by Cursor when needed. You can verify it's working by:

1. Opening Cursor
2. The memory server will be started automatically when Cursor launches
3. You can check the server status by visiting http://localhost:3333/mcp/v1/health

Available MCP Endpoints

SSE Connection

• Endpoint : GET /mcp/v1/sse
• Query Parameters :
  • subscribe: Comma-separated list of events to subscribe to (optional)
• Events :
  • connected: Sent on initial connection
  • memory.created: Sent when new memories are created
  • memory.updated: Sent when existing memories are updated

Memory Operations

1. Create Memory

POST /mcp/v1/memory
Content-Type: application/json

{
  "type": "learning",
  "content": {
    "topic": "Express.js",
    "details": "Express.js is a web application framework for Node.js"
  },
  "source": "documentation",
  "tags": ["nodejs", "web-framework"],
  "confidence": 0.95
}

2. Search Memories

GET /mcp/v1/memory/search?query=web+frameworks&type=learning&tags=nodejs

3. List Memories

GET /mcp/v1/memory?type=learning&tags=nodejs,web-framework

Health Check

GET /mcp/v1/health

Response Format

All API responses follow the standard MCP format:

{
  "status": "success",
  "data": {
    // Response data
  }
}

Or for errors:

{
  "status": "error",
  "error": "Error message"
}

Memory Schema

• id: Unique identifier
• type: Type of memory (learning, experience, etc.)
• content: Actual memory content (JSON)
• source: Where the memory came from
• embedding: Vector representation of the content (384 dimensions)
• tags: Array of relevant tags
• confidence: Confidence score (0-1)
• createdAt: When the memory was created
• updatedAt: When the memory was last updated