概要
What is Agent Memory?
Agent Memory is an MCP server that exposes an agent-memory-daemon to any MCP-compatible client. The server acts as a thin filesystem bridge so that agents can read, append, and search memory through the Model Context Protocol, while the daemon handles consolidation and extraction.
How to use Agent Memory?
—
Key features of Agent Memory
- Exposes agent-memory-daemon via the Model Context Protocol
- Acts as a lightweight filesystem bridge for memory operations
- Supports reading, appending, and searching agent memory
- Compatible with Kiro, Claude Desktop, Cursor, and other MCP clients
- Separates memory storage (server) from consolidation and extraction (daemon)
Use cases of Agent Memory
- Allow AI agents to persist and retrieve memories across different sessions
- Enable memory search functionality for any MCP-compatible client
- Decouple memory management logic from client-specific implementations
FAQ from Agent Memory
Which clients does Agent Memory work with?
Agent Memory works with any MCP-compatible client, including Kiro (CLI & IDE), Claude Desktop, Cursor, and others.
What is the difference between the server and the daemon?
The agent-memory-daemon handles the thinking—consolidation and extraction—while the server is a thin filesystem bridge that enables memory operations through MCP.
What memory operations does Agent Memory support?
The server enables agents to read, append, and search memory through the Model Context Protocol.
What are the runtime dependencies of Agent Memory?
—
