turbyho/mem-context

What it does

turbyho/mem-context is a Model Context Protocol (MCP) server that provides intelligent temporal memory management for AI applications. It uses LanceDB vector search combined with a 6-factor relevance scoring system to automatically capture, organize, and retrieve conversation history. The tool implements weight-decay scoring to prioritize recent memories while maintaining access to older context, and includes LLM-powered consolidation to summarize and deduplicate stored information.

How to set up

Installation is straightforward via pip: pip install memcontext. Once installed, the MCP server can be configured to connect with compatible AI platforms and LLM applications. The server automatically captures sessions and requires minimal configuration to begin storing and retrieving temporal context.

Key features

• Automatic session capture and memory persistence
• LanceDB-powered vector search for semantic retrieval
• 6-factor relevance scoring algorithm
• Weight-decay scoring for temporal memory prioritization
• LLM-based consolidation and deduplication
• Cross-platform MCP server compatibility
• Python-based implementation with pip installation