Vector search · facts read November 7, 2026
pgvector (in Postgres)
An open-source Postgres extension that stores and searches embeddings next to your other data. No extra vendor: vectors live in the Postgres you already have, so tenant filters are a plain WHERE clause.
- Free plan
- Free: on Neon it is "available on every Neon plan with no add-on or paid tier required"; you pay only for database storage and compute.
- Paid from
- Not stated on the pricing page
- Monthly cost at 1,000 users
- $0/month Included in your Postgres; ~50,000 chunks of 1,024-dim vectors is roughly 0.2 GB of storage.
- Free plan allows commercial use
- Not stated
- Open source
- Yes
- Self-hostable
- Yes
Source: pgvector (in Postgres) pricing, read November 7, 2026. Prices change; check the vendor's page before you commit.
Best for
- Up to a few million chunks
- Filtering vector search by tenant with SQL
- Fewest moving parts
Watch out for
- Its HNSW index supports vectors up to 2,000 dimensions (halfvec up to 4,000); pick an embedding size that fits.
- Vectors count toward your database storage (Neon Free: 1 GB per project).
How Cuely scores pgvector (in Postgres)
The published scores, from −3 to +3, that decide when Cuely picks it. Method
- +3fewer vendors. Vectors stay in your existing Postgres; no new vendor or sync job.
- +3keep costs as low as possible. No separate bill; it uses the database you already pay for.
- +3prefers open source or self-hosting. Open-source extension that runs on any Postgres.
- −1high volume. Past tens of millions of vectors, a dedicated vector store is easier to run.