Mem0 MCP Server - Node.js (HTTP-Only)

What is Mem0 MCP Server - Node.js (HTTP-Only)?

A Node.js implementation of the Mem0 MCP server that runs exclusively via HTTP API for simplified deployment and integration. It provides tools to save, retrieve, and search memories locally in the active project directory, and is designed for use with MCP clients like VS Code Roo Code and Claude Desktop.

How to use Mem0 MCP Server - Node.js (HTTP-Only)?

Install dependencies with npm install, copy .env.example to .env, configure environment variables, then start the server with npm start (default port 8484). For specific workspaces, use npm run start:workspace /path/to/your/project. Integrate with MCP clients by configuring the appropriate transport (stdio, SSE, or HTTP) and providing the API token in environment or configuration.

Key features of Mem0 MCP Server - Node.js (HTTP-Only)

• HTTP-only operation for simplified deployment
• Workspace integration with automatic memory storage
• Token-based authentication and rate limiting
• Local file storage in a .Mem0-Files directory
• Dual transport support (stdio, SSE, and HTTP)
• Docker support for containerized deployment

Use cases of Mem0 MCP Server - Node.js (HTTP-Only)

• Saving project context and coding patterns in VS Code
• Searching stored memories to recall past decisions
• Retrieving all stored memories for a specific user
• Integrating persistent memory into AI assistants like Claude Desktop
• Self-hosting a memory backend for MCP-compatible applications

FAQ from Mem0 MCP Server - Node.js (HTTP-Only)

What transports does the server support?

The server supports stdio, SSE, and HTTP transports. HTTP is the primary mode and is enabled by default. SSE transport requires the server to be started with TRANSPORT=sse HTTP_SERVER_ENABLED=true npm start.

How are memories stored?

Memories are stored locally in JSON files within a .Mem0-Files directory in the workspace root. Each memory file contains the original text, metadata (created/updated timestamps), and a unique UUID. An index file maps user IDs to their memories.

How do I secure the API?

Set the API_TOKEN environment variable to a secure token. All HTTP API requests must include an Authorization: Bearer <token> header. Rate limiting is also enabled by default (100 requests per 15-minute window).

Can I run this as a Docker container?

Yes. Use npm run docker:build to build the Docker image and npm run docker:run to start a container. The container exposes the HTTP API on port 8484.

Is there any LLM integration?

Configuration options for LLM providers (OpenAI, OpenRouter, Ollama) are available in the .env file, but full integration for memory processing and semantic search is planned for future releases.