Mem Context

Local MCP server that gives an assistant a temporal memory store backed by LanceDB vector search, weight-decay scoring and LLM consolidation.

What you can do

• Stores and retrieves memories for the assistant using LanceDB vector search.
• Ranks recalled memories with weight-decay scoring so older or less relevant items fade over time.
• Consolidates stored memories through an LLM.
• Runs as a local stdio process on the user's machine.

When to use it

Worth adding when you want an assistant to keep context across sessions and recall earlier information with recency-aware ranking, rather than starting from scratch each time.

Requirements

Runs locally as a stdio process, installed from the PyPI package mem-context, so a Python environment is needed. No authentication, environment variables or headers are declared. An LLM is involved in consolidation, so a model or provider must be available to it.