supermemory
Long-term memory layer that follows your AI across tools
gbrain is Garry Tan's organized long-term memory system for AI agents, storing people, companies, and timelines so an assistant can retrieve context. It is aimed at operators who manage many founders, clients, or contacts.
It uses Postgres as the database with embeddings and BM25 ranking for retrieval, organizing entities and their timelines. It is accessed through an AI assistant or Claude-style interface to give agents structured memory.
Cut LLM cost by 95%, replace ElevenLabs, and 10 top GitHub repos
Long-term memory layer that follows your AI across tools
Convert PDFs, Office files, audio and YouTube into clean Markdown
Self-evolving context database giving agents organized, searchable memory
Visual web search for agents using screenshots, not just text
Persistent memory that carries agent context between sessions
Compress tool outputs, logs and RAG chunks before the LLM sees them