Agent♥︎Age
Catalog

MemoCat — Shared AI Memory

Official

by MontyGovernance · Python

One shared, persistent memory across AI agents and systems, with semantic recall and live updates.

io.github.MontyGovernance/memocat-mcp (MCP)

This self-hosted Model Context Protocol (MCP) server provides one shared, persistent memory across AI agents and systems, with semantic recall and live updates. MCP clients can write to the memory and read each other’s data, supporting long-term, searchable information for multi-agent setups.

🛠️ Key Features

  • One shared, persistent memory across AI agents and systems
  • Semantic recall
  • Live updates
  • Semantic-search backed by embeddings/vector concepts
  • MCP server for model-context-protocol clients

🚀 Use Cases

  • Enable Claude, Codex, Cursor, and other MCP clients to share memory
  • Multi-agent workflows requiring shared state
  • Retrieval of prior context via searchable, semantic memory
  • Self-hosted deployments for long-term memory needs

⚡ Developer Benefits

  • Cross-agent memory read/write via MCP
  • Semantic-search and vector-search oriented memory access
  • Consistent long-term-memory across tools and systems

⚠️ Limitations

  • Functionality described only as shared persistent memory with semantic recall and live updates; no further constraints are specified.

Topics

agent-memoryai-memoryclaudeclaude-codecodexcursorembeddingsllmlong-term-memorymcpmcp-servermodel-context-protocolmulti-agentpersistent-memoryragself-hostedsemantic-searchshared-memoryvector-databasevector-search