agentmem

Agentmem is a lightweight, production-ready persistent memory layer for AI agents, enabling them to store and recall long-term information across sessions using a single SQLite file. It offers hybrid keyword and semantic search, multi-tiered memory, namespaces for isolation, and advanced features like auto-entity extraction, fact versioning, and memory consolidation. Designed for high speed, minimal dependencies (works even in pure Python, no cloud), and direct integration into agentic and Model Context Protocol (MCP) workflows, it is ideal for developers building autonomous or context-aware AI that needs robust, scalable memory.