michael-denyer/memory-mcp

Overview

A Python-based MCP server that provides memory capabilities inspired by the Engram system. It uses a hot cache for fast retrieval and pattern mining to extract recurring themes from stored data. Designed to give AI agents persistent context across sessions.

Best for

Best for
Developers prototyping memory-enhanced AI agents in Python

Use cases

• Adding long-term memory to chatbot agents
• Caching conversation history for low-latency recall
• Mining interaction patterns to personalize responses

Notes

A Python-based MCP server that provides memory capabilities inspired by the Engram system. It uses a hot cache for fast retrieval and pattern mining to extract recurring themes from stored data. Designed to give AI agents persistent context across sessions.

6 stars on GitHub. Last updated 2026-01-25. Licensed MIT.

Use cases

• Adding long-term memory to chatbot agents
• Caching conversation history for low-latency recall
• Mining interaction patterns to personalize responses

Pros

• Lightweight Python implementation easy to integrate
• Hot cache reduces latency for frequently accessed memories
• Pattern mining adds contextual awareness beyond simple storage

Cons

• Very early stage with only 6 GitHub stars and limited community
• Documentation and examples are sparse
• Pattern mining may introduce overhead for simple use cases

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.