Memorious MCP

What is Memorious MCP?

Memorious MCP is a local and persistent semantic memory server for AI assistants using the Model Context Protocol (MCP). It uses ChromaDB for vector similarity search and FastMCP for efficient tool implementation, and runs entirely on your machine—no data ever leaves it.

How to use Memorious MCP?

Install via uvx memorious-mcp or set up in a virtual environment with uv sync and uv run memorious-mcp. Add the server to your MCP client configuration with the command uvx and arguments ["memorious-mcp"]. Use the three built-in tools: store, recall, and forget.

Key features of Memorious MCP

• 100% local and private – all processing stays on your machine
• Persistent memory across sessions using ChromaDB disk-based storage
• Semantic search via vector embeddings for similarity-based retrieval
• Simple three-tool API for memory management
• FastMCP integration for efficient MCP server implementation
• Canonical key design optimized for short, embedding-friendly keys
• No external API calls or cloud dependencies required

Use cases of Memorious MCP

• Personal assistant memory – remember user preferences, habits, and personal information
• Context preservation – maintain conversation context across sessions
• Knowledge management – store and retrieve project-specific information
• Personalization – enable AI assistants to provide tailored responses based on stored preferences
• Privacy‑first AI – keep sensitive personal data local while still having persistent memory

FAQ from Memorious MCP

How does Memorious MCP ensure privacy?

All data processing and storage happens locally on your machine—no network requests, no cloud dependencies, and no data ever leaves your computer.

What are the system requirements?

Python version 3.12 or higher is required. The server uses ChromaDB and FastMCP, which are installed automatically.

Where is my data stored?

Data is stored in the ./.memorious directory (configurable) using ChromaDB’s persistent disk-based storage.