Back to Ecosystem Pulse
ProtocolProduction ReadyMCP
mcp-memory-service
by doobidoo
MCP-compatible persistent memory service with semantic search and consolidation
Python
Updated May 30, 2026
Share:
Overview
Provides a persistent memory service for agent pipelines exposing a REST APIREST API and knowledge-graph backed vector store. Implements the Model Context Protocol (MCP)Model Context Protocol (MCP) for long-term memory, semantic search, and autonomous consolidation of memories. Designed to plug into LangGraph, CrewAI, AutoGen and Claude-based systems with SQLite/vector DB support for inexpensive deployment.
Why It Matters
As agent networks run over weeks and months, persistent, queryable memories become the backbone of agent-to-agent trust and accountability. By centralizing memory with MCP compatibility and a knowledge-graph layer, teams can reconstruct agent histories, validate past claims, and measure agent track records. That historical signal is critical for A2A evaluation and continuous reputation-aware decisioning. This supports A2A evaluation and interoperability across agents.
Best For
Teams building multi-agent pipelines (LangGraph, CrewAI, AutoGen, Claude) that need long-term, queryable memory and traceable agent histories.
Applications
- Persisting agent observations and decisions to reconstruct agent behavior for audits
- Providing semantic retrieval and RAG context to agents across sessions
- Autonomously consolidating redundant memories to reduce noise and improve memory quality
Works With
autogencrewaianthropic
Topics
agent-memoryagentic-aiai-agentsautogenclaudecrewaiknowledge-graphlanggraphlong-term-memorymcp+10 more
Similar Tools
langchainautogen
Keywords
agent track recordmcpmulti-agent trustlong-term-memory