Agentflow: open-source collective memory network for agents via FastMCP and pgvector
agentflow0328 · reddit · 2026-10-16
The author open-sources Agentflow, a cross-agent collective memory network tackling "epistemic isolation"—agents discarding hard-won debugging trajectories when context resets. Agents query shared, machine-verified remedies via a searchplaybooks FastMCP tool (pgvector cosine similarity on Supabase) and submit verified fixes back with submitremedy.
Architecture highlights
- FastMCP server compatible with Cursor, Claude Desktop, and custom Python agent runtimes
- Background vectorization queue via Postgres FOR UPDATE SKIP LOCKED, no Celery/Redis needed
- Dual embeddings: OpenAI in production, deterministic subword-hash fallback for fully offline CI
- Hardening: SHA-256 token hashing, payload limits (20KB traces / 100KB playbooks), rate limiting
Open source at github.com/adithyasaineeli/agentflow.
More from coding & agent
- Sakana AI ships Agentic Scientist, orchestrating coding agents for automated scientific discovery — hardmaru · 2026-10-16
- One trailing dot bypassed VS Copilot's path check, enabling RCE (CVE-2026-103572) — wunderwuzzi23 · 2026-10-16
- Celestine MCP server adds astrology charts and transits to AI agents — modelcontextprotocol · 2026-10-16
- ProfitPlay MCP Server lets AI agents trade crypto and stock prediction markets — modelcontextprotocol · 2026-10-16
- Polyhaven flagged as a go-to resource for agents building realistic 3D scenes — jason_mayes · 2026-10-16
- Excaliflow goes fully open-source: hand-drawn diagrams with live animated data flows in a single HTML file — Saboo_Shubham_ · 2026-10-16