GBrain
Garry Tan's open-source agent brain: shared memory with sources, corrections and a knowledge graph, served to Claude, Codex, OpenClaw, Hermes and others over MCP.
Tags
Description
GBrain stores explicit facts with provenance, supports corrections and withdrawal, and makes the same memory available across agents. Start keyless with keyword retrieval, then add semantic search, a cited synthesis layer (answers with an explicit note on what the brain doesn't know), a typed knowledge graph and background enrichment/consolidation jobs. Connects to coding agents and personal agents over MCP; gbrain mcp expose publishes a local brain on a Tailscale tailnet (with --funnel for cloud agents). Install requires Bun ≥1.4 via bun install -g github:garrytan/gbrain; the README warns that the npm package named gbrain is unrelated. A hosted team version is offered at gbrain.io. Ships llms.txt, llms-full.txt and AGENTS.md for agents.
Agent access
Related
- memU: Agent-driven personal memory stored as a shared LLM wiki across sessions, agents and devices; distills reusable skills from Codex, Claude Code, Cursor and OpenClaw history.
- Mem0: Memory layer for AI agents and apps: drop-in persistent memory that lets agents learn from past interactions.
- Basic Memory: Local-first MCP server that gives AI assistants persistent memory stored as Markdown files in a knowledge base you own.
- Honcho: Open-source memory infrastructure for stateful agents that models people, agents, groups and projects over time; managed or self-hosted.
- OpenClaw: Open-source personal AI assistant that runs on your own machine behind one Gateway and talks to you over WhatsApp, Telegram, Slack, Discord, iMessage and 20+ other channels.
- Hermes Agent: Nous Research's self-improving open-source agent: terminal UI plus messaging gateway, agent-curated memory and skills it writes and refines from experience.
Sources
Machine-readable
- JSON
- Markdown
- Source file on GitHub (edit via pull request)