Knowledge brain for OpenClaw agents. Gives agents searchable, indexed memory over markdown repos using Postgres + pgvector + hybrid search. Paste install instructions into any OpenClaw agent and it handles the rest.
Reference Numbers (from one user, 6 months)
- 10,000+ markdown files indexed and searchable
- 3,000+ people with compiled dossiers and relationship history
- 13 years of calendar data (21,000+ events)
- 5,800+ Apple Notes going back to 2009
Why This Matters
Directly relevant to the Obsidian vault + Claude Code setup. Current memory system is file-based (MEMORY.md + individual memory files). GBrain offers vector search over the full vault — could surface connections and context that flat-file grep misses. The “people dossiers + relationship history” pattern is exactly what the People folder is building toward.
Questions
- How does it compare to just using Claude Code’s native Grep/Glob over the vault?
- What’s the latency on hybrid search vs. ripgrep?
- Does it work with Claude Code or only OpenClaw?
- Self-hosted Postgres requirement — overhead worth it at current vault size (~2K files)?
Source
- Platform: Instagram
- Author: Garry Tan (@garrytan, verified)
- Date: 2026-04-12
- Tagline: “The memex Vannevar Bush imagined, built for people who think for a living”
part of tooling