8 engineers on Hacker News describe using pgvector first-hand: 5 endorse it, 1 regret it, 0 are mixed.
“Instead we run everything on TigerData (creators of TimescaleDB)’s PostgreSQL platform (TimescaleDB + pgVector + pgAI).”
endorse · production · ishita159 · 2025-08-19 · source ↗ · wrong label?“once we switched to PGVector in Django the need for a lot of things went away.”
endorse · production · stuaxo · 2025-05-11 · source ↗ · wrong label?“Everything is on Postgres and pgvector, so you can do all the Postgres stuff and decrease corpus size by document metadata filtering”
endorse · production · jonathan-adly · 2024-12-05 · source ↗ · wrong label?“I've been using pgvector, it has worked as expected, which is all I want.”
endorse · personal · clwg · 2023-12-12 · source ↗ · wrong label?“which is perhaps why pgvector suffers more under higher load.”
regret · production · icyfox · 2023-11-26 · source ↗ · wrong label?“we're[0] using pgvector via supabase in prod and we've had 0 issues.”
endorse · production · mritchie712 · 2023-06-08 · source ↗ · wrong label?“while keeping recall latency low using pgvector.”
neutral · production · tylerrecall · 2025-12-17 · source ↗ · wrong label?“Yes, all use PGvector. In our hosting all instances default to having the vector features enabled”
neutral · production · xfalcox · 2025-11-04 · source ↗ · wrong label?