agentmemory
Persistent memory for AI coding agents that runs locally
Visit agent-memory.dev ↗What is agentmemory?
agentmemory gives AI coding agents persistent memory across sessions, allowing them to recall context without re-explaining work. It runs entirely on your machine with zero external databases, working with any MCP client including Claude Code, Cursor, GitHub Copilot CLI, and others. The project achieves 95.2% retrieval recall on LongMemEval-S while using 92% fewer input tokens than full-context replay.
The system captures every tool call, prompt, and agent action through 12 auto-hooks, then uses hybrid retrieval combining BM25, vector search, and knowledge graph ranking to recall relevant memories. It includes 54 MCP tools and 134 REST endpoints for integration, with optional LLM-powered consolidation that compresses observations into semantic memories and removes duplicates. Built on the iii engine, the entire runtime operates as a single local process.
Momentum
Daily activity over the last 30 days: upvotes, unique visits, new GitHub stars and Hacker News points.
How do we know it's vibe coded?
vibetrend discovered agentmemory automatically. Evidence: GitHub rohitg00/agentmemory; repo has AGENTS.md. View source.
Similar vibe-coded apps
Turn your codebase into a queryable knowledge graph for AI coding assistants
Operating layer for AI coding agents with skills, memory, planning and security
Self-hosted AI agent with persistent memory and multi-platform messaging
Open-source AI platform for short-form video and animated series creation
Open-source canvas workbench for AI-powered image and video creation
Community marketplace for DeepSeek Harness Web themes, pets, plugins
Open-source, AI-native knowledge base with Evernote-style interface