MARM
Stop re-explaining yourself to your AI. MARM gives every session persistent memory, cross-agent context sharing, and structured reasoning that builds as you work, not resets. Includes a live web dashboard for browsing memories, logs, and sessions wit
What is MARM?
MARM is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to stop re-explaining yourself to your ai. marm gives every session persistent memory, cross-agent context sharing, and structured reasoning that builds as you work, not resets. includes a live web dashb...
Stop re-explaining yourself to your AI. MARM gives every session persistent memory, cross-agent context sharing, and structured reasoning that builds as you work, not resets. Includes a live web dashboard for browsing memories, logs, and sessions wit
This server falls under the Knowledge & Memory category on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- Stop re-explaining yourself to your AI. MARM gives every ses
Use Cases
Maintainer
Works with
Installation
PIP
pip install marm-mcp-serverManual Installation
pip install marm-mcp-serverConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use MARM
MARM (Memory-Augmented Reasoning Module) is a persistent memory MCP server for AI assistants that eliminates the need to re-explain context at the start of every session. It stores memories using semantic embeddings, organizes them into named sessions and notebooks, and exposes nine dedicated tools so agents can recall past decisions, log ongoing work, and share context across multiple agents. Developers who run long-running projects with Claude Code or other MCP clients will find MARM prevents context loss across restarts.
Prerequisites
- Python 3.10+ and pip, or Docker installed
- An MCP-compatible client such as Claude Desktop or Claude Code
- A generated MARM API key (required for HTTP/Docker deployment)
- Optional: Ollama running locally if you want local embedding support
Install the MARM MCP server
Install via pip for stdio mode, or pull the Docker image for HTTP mode with multi-agent support.
pip install marm-mcp-server
# OR for Docker:
docker pull lyellr88/marm-mcp-server:latestGenerate an API key (Docker/HTTP mode only)
If running MARM as an HTTP server over Docker, generate an authentication key first.
docker run --rm lyellr88/marm-mcp-server:latest --generate-keyStart the Docker container
Run the container exposing port 8001, binding the data volume to ~/.marm, and providing the generated API key.
docker run -d --name marm-mcp-server \
-p 127.0.0.1:8001:8001 \
-e SERVER_HOST=0.0.0.0 \
-e MARM_API_KEY=your-generated-key \
-v ~/.marm:/home/marm/.marm \
lyellr88/marm-mcp-server:latestAdd MARM to your MCP client (stdio mode)
For local stdio use, register the marm-mcp-stdio command with Claude Code or your MCP client.
claude mcp add --transport stdio marm-memory-stdio marm-mcp-stdioAdd MARM to your MCP client (HTTP mode)
For the HTTP Docker deployment, register the local endpoint with an Authorization header.
claude mcp add --transport http marm-memory http://localhost:8001/mcp \
--header "Authorization: Bearer your-generated-key"Verify the connection
Ask the agent to log a session entry. MARM will respond confirming the memory was stored without requiring any manual tool invocation.
MARM Examples
Client configuration (stdio mode)
Claude Desktop config for local stdio mode installation via pip.
{
"mcpServers": {
"marm-memory": {
"command": "marm-mcp-stdio",
"args": []
}
}
}Client configuration (HTTP/Docker mode)
Claude Desktop config for the Docker HTTP deployment with Bearer token authentication.
{
"mcpServers": {
"marm-memory": {
"url": "http://localhost:8001/mcp",
"headers": {
"Authorization": "Bearer your-generated-key"
}
}
}
}Prompts to try
MARM tools are invoked automatically by natural language — no need to call tools manually.
- "Log this session as 'Project Alpha' and save this conversation as 'database design discussion'"
- "Search your memory for what we discussed about authentication last week"
- "Remember this code snippet in your notebook for later use"
- "Show me all entries in the 'Project Alpha' session"
- "Generate a summary of everything we've worked on this sprint"Troubleshooting MARM
Docker container starts but the MCP client cannot connect to port 8001
Ensure you used -p 127.0.0.1:8001:8001 and SERVER_HOST=0.0.0.0. Confirm with: docker logs marm-mcp-server and curl http://localhost:8001/mcp.
Authentication errors when connecting via HTTP mode
Regenerate the API key with: docker run --rm lyellr88/marm-mcp-server:latest --generate-key, update the MARM_API_KEY environment variable in the container, and update the Authorization header in your MCP client config.
Memories are not persisting between Docker restarts
Ensure the volume mount -v ~/.marm:/home/marm/.marm is present in the docker run command. Without it, data is stored only in the ephemeral container layer.
Frequently Asked Questions about MARM
What is MARM?
MARM is a Model Context Protocol (MCP) server that stop re-explaining yourself to your ai. marm gives every session persistent memory, cross-agent context sharing, and structured reasoning that builds as you work, not resets. includes a live web dashboard for browsing memories, logs, and sessions wit It connects AI assistants to external tools and data sources through a standardized interface.
How do I install MARM?
Install via pip with: pip install marm-mcp-server. Then configure your AI client to connect to this MCP server.
Which AI clients work with MARM?
MARM works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is MARM free to use?
Yes, MARM is open source and available under the MIT License license. You can use it freely in both personal and commercial projects.
MARM Alternatives — Similar Knowledge & Memory Servers
Looking for alternatives to MARM? Here are other popular knowledge & memory servers you can use with Claude, Cursor, and VS Code.
MemPalace
★ 52.6kA local AI memory system that stores all conversations verbatim and organizes them into navigable structures. It provides 19 MCP tools for AI assistants to search and retrieve past decisions, debugging sessions, and architecture debates automatically
Kratos
★ 25.7k🏛️ Memory System for AI Coding Tools - Never explain your codebase again. MCP server with perfect project isolation, 95.8% context accuracy, and the Four Pillars Framework.
Context Mode
★ 15.4kAn MCP server that preserves LLM context by intercepting large data outputs and returning only concise summaries or relevant sections. It enables efficient sandboxed code execution, file processing, and documentation indexing across multiple programm
Memu
★ 13.7kMemory for 24/7 proactive agents like OpenClaw.
MemOS
★ 9.3kMemOS (Memory Operating System) is a memory management operating system designed for AI applications. Its goal is: to enable your AI system to have long-term memory like a human, not only remembering what users have said but also actively invoking, u
Everos
★ 5.4kBuild, evaluate, and integrate long-term memory for self-evolving agents.
Browse More Knowledge & Memory MCP Servers
Explore all knowledge & memory servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.
Set Up MARM in Your Editor
Choose your AI client for step-by-step setup instructions.
Quick Config Preview
Add this to your claude_desktop_config.json or .cursor/mcp.json
Ready to use MARM?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.