gbrain

Tech stack & purpose

GBrain is a personal knowledge brain and agent memory system built to help LLM agents and humans organize, retrieve, and reason over information. It's designed for both individual knowledge bases and team-scale deployments, with a focus on retrieval quality and hybrid search capabilities. The project is written in TypeScript and uses Bun as its build tool and package manager, with a pluggable database engine architecture supporting either PGLite (an embedded Postgres via WebAssembly requiring no configuration) or a managed Postgres instance with pgvector for vector search. Created by Garry Tan, GBrain integrates with multiple agent platforms including Claude, Codex, Cursor, and others through an MCP (Model Context Protocol) interface, and provides both a CLI and programmatic API for managing knowledge across configurable brains (databases) and sources (repositories within those databases).

Languages

Contributors (top 30 of 130)

• Garry Tan456 commits
• Time Attakc164 commits
• Masa54 commits
• garrytan-agents18 commits
• Ziyang Guo10 commits
• Javier Aldape9 commits
• Paolo Belcastro8 commits
• Sean Gearin6 commits
• paul-03206 commits
• Sina Matian6 commits
• Anton Senkovskiy5 commits
• zsimovanforgeops5 commits
• cybernaut64045 commits
• alexey-metaengage5 commits
• Brett5 commits
• spiky02plateau5 commits
• mzkarami4 commits
• caterpillarC154 commits
• test4 commits
• Wintermute4 commits
• Elliot Drel4 commits
• Ryan Ayers4 commits
• 
  bender23523 commits · 0 followers
• Sailesh Sivakumar3 commits
• arisgysel-design3 commits
• maxpetrusenkoagent3 commits
• Benjamin D. Smith3 commits
• daragao33 commits
• johnnymn3monic2 commits
• Mikhail Merkulov2 commits