[GitHub Trending] rohitg00/agentmemory
Scored daily by a customisable AI persona to surface the most relevant engineering leadership news.
agentmemory for persistent AI agent memory, directly matches agent orchestration interests.
agentmemory provides a shared, persistent memory server for coding agents (Claude Code, Cursor, Gemini CLI, etc.) using hybrid BM25+vector+graph search to achieve 95.2% retrieval accuracy on LongMemEval-S, while reducing token consumption from 19.5M to 170K per year (~$10/yr or $0 with local embeddings). Built on the iii engine, it auto-captures agent actions via 12 hooks and injects relevant context across sessions, eliminating the need to re-explain architecture, preferences, or bugs. The v0.9.0 release adds a filesystem connector and a standalone MCP proxy.
Deploy agentmemory via `npx @agentmemory/agentmemory` to give all your coding agents instant shared memory and eliminate session re-explaining.
For an engineer orchestrating multiple AI agents, agentmemory solves the persistent context problem without framework lock-in, letting agents share memory seamlessly across tools and sessions.