local-memory-mcp

What it does

local-memory-mcp provides a sophisticated persistent memory layer for Claude, Cursor, and Codex through the Model Context Protocol. It enables AI assistants to store, retrieve, and reason over information across conversations using a hybrid retrieval system combining BM25 keyword search and vector similarity ranking. The tool includes advanced features like bi-temporal queries for tracking information changes over time and automated contradiction detection to maintain knowledge consistency.

How to set up

Installation is simple via npm: npx @studiomeyer/local-memory-mcp. The tool runs entirely locally on your machine without requiring cloud services or external API keys. It uses SQLite for data persistence and automatically manages a knowledge graph structure. Once installed, it integrates directly with Claude, Cursor, or Codex as an MCP server.

Key features

• 21 integrated MCP tools for memory operations
• Hybrid retrieval combining BM25 and vector RRF (Reciprocal Rank Fusion)
• Bi-temporal query support for historical reasoning
• Automatic contradiction detection across knowledge base
• SQLite database with knowledge graph structure
• Completely local and privacy-preserving (no cloud)
• TypeScript implementation