Memomind
Give your AI agent a brain that remembers. Local memory system for Claude Code — 100% private, GPU-accelerated, zero cloud dependency.
What is Memomind?
Memomind is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to give your ai agent a brain that remembers. local memory system for claude code — 100% private, gpu-accelerated, zero cloud dependency.
Give your AI agent a brain that remembers. Local memory system for Claude Code — 100% private, GPU-accelerated, zero cloud dependency.
This server falls under the Knowledge & Memory category on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- Give your AI agent a brain that remembers. Local memory syst
Use Cases
Maintainer
Works with
Installation
Manual Installation
npx memomindConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use Memomind
MemoMind is a fully local, GPU-accelerated persistent memory system for AI coding agents that uses PostgreSQL with pgvector and a knowledge graph to give agents structured, evolving memory across sessions. It exposes three MCP tools — retain (store facts), recall (4-way hybrid search combining semantic, BM25, knowledge graph, and temporal retrieval), and reflect (deep cross-memory reasoning) — so Claude Code and other MCP agents can autonomously remember preferences, architectural decisions, and project context without any manual effort from the developer.
Prerequisites
- Python 3.11+ installed on Windows, Linux, or WSL2
- PostgreSQL 17 with the pgvector extension (the install script handles this on Linux)
- An NVIDIA GPU with CUDA 12.4+ is recommended for fast local embeddings (CPU fallback is available)
- An LLM API key from any OpenAI-compatible provider (OpenAI, Anthropic, DeepSeek, Groq, Ollama, etc.) for fact extraction
- Claude Code or another MCP-compatible client with stdio transport support
Clone the repository
Download the MemoMind source code to your machine.
git clone https://github.com/24kchengYe/MemoMind.git
cd MemoMindInstall on Linux or WSL2 (automated)
Run the provided install script which sets up Python dependencies, PostgreSQL 17, pgvector, and the systemd service automatically.
sudo bash install.shConfigure your LLM API key
Edit serve.py to set your LLM provider credentials. MemoMind only uses the LLM for fact extraction (not chat), so a cheap fast model like deepseek-chat or gpt-4o-mini works well.
# In serve.py, set these three variables:
LLM_API_KEY = "your-api-key"
LLM_BASE_URL = "https://api.openai.com/v1" # or your provider's URL
LLM_MODEL = "gpt-4o-mini"Start the MemoMind service
On Linux/WSL2, start the service with systemd. The memory API will be available at http://localhost:19999 and the web dashboard at http://localhost:9999.
# Linux/WSL2 with systemd:
sudo systemctl start memomind
# Verify the service is healthy:
curl http://localhost:19999/healthRegister MemoMind as an MCP server in Claude Code
Add MemoMind to Claude Code using the MCP add command. Use the appropriate command for your platform (Linux native or WSL2 from Windows).
# Linux native:
claude mcp add --scope user --transport stdio memomind \
-- /opt/memomind-env/bin/python3 /opt/memomind-env/mcp_stdio.py
# From Windows using WSL2:
claude mcp add --scope user --transport stdio memomind \
-- wsl -d Ubuntu -u memomind -e //opt/memomind-env/mcp-entry.shExplore memories via the web dashboard
Open the MemoMind dashboard to browse, search, and manage all stored memories. The dashboard supports knowledge graph visualization, timeline view, and dual keyword/semantic search.
open http://localhost:9999Memomind Examples
Client configuration
MemoMind registers via the claude mcp add CLI command rather than a JSON config block. Once registered, it automatically activates its retain/recall/reflect tools in every Claude Code session.
# Run this once to register:
claude mcp add --scope user --transport stdio memomind \
-- /opt/memomind-env/bin/python3 /opt/memomind-env/mcp_stdio.py
# Verify registration:
claude mcp listPrompts to try
After MemoMind is connected, Claude Code autonomously calls retain and recall. You can also prompt explicit memory operations.
- "Remember that we decided to use FastAPI instead of Flask for all new microservices in this project."
- "What do you remember about our database schema decisions from previous sessions?"
- "Reflect on all the debugging sessions we've had and identify the most common error patterns in this codebase."Troubleshooting Memomind
Health check returns database connection error
Verify PostgreSQL is running with `systemctl status postgresql` and that the pgvector extension is installed. On Linux, re-run `sudo bash install.sh` to repair the database setup. Check logs with `journalctl -u memomind`.
Embeddings are very slow without GPU acceleration
MemoMind uses the bge-m3 model which benefits greatly from a CUDA GPU. On CPU-only systems, keyword search (20ms) still works fast but semantic recall will be slower (several seconds per query). Install CUDA 12.4+ drivers to enable GPU acceleration.
Claude Code does not seem to be calling retain or recall automatically
MemoMind relies on Claude Code autonomously deciding when to use the tools. Add an instruction to your CLAUDE.md file: 'Use the memomind retain tool to store important decisions and recall tool to check memory before answering questions about this project.' This primes the model to use memory tools consistently.
Frequently Asked Questions about Memomind
What is Memomind?
Memomind is a Model Context Protocol (MCP) server that give your ai agent a brain that remembers. local memory system for claude code — 100% private, gpu-accelerated, zero cloud dependency. It connects AI assistants to external tools and data sources through a standardized interface.
How do I install Memomind?
Follow the installation instructions on the Memomind GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.
Which AI clients work with Memomind?
Memomind works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is Memomind free to use?
Yes, Memomind is open source and available under the MIT license. You can use it freely in both personal and commercial projects.
Memomind Alternatives — Similar Knowledge & Memory Servers
Looking for alternatives to Memomind? 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 Memomind 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 Memomind?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.