pgvector
pgvector is an open-source PostgreSQL extension that adds vector similarity search to any PostgreSQL database. Free and self-hosted. Handles the majority of RAG use cases below 1 million vectors without dedicated infrastructure. The lowest-friction path to vector search for teams already on PostgreSQL.
pgvector handles 80% of production RAG workloads at zero marginal cost if you're already on PostgreSQL - start here and only migrate when you hit actual performance limits. Purpose-built vector databases like Qdrant or Pinecone make sense when you exceed 10M vectors or need advanced filtering performance that pgvector can't deliver.
· Expert analysis by Oleh Kem, Founder, ComparEdge
Highly rated vector databases software built for Teams with existing PostgreSQL infrastructure wanting to add vector search with zero new services who need focused workflows.
Top Pros
- No new infrastructure — runs inside existing PostgreSQL
- SQL interface familiar to every developer
- ACID transactions across vectors and relational data
Watch Out For
- Performance trails purpose-built vector DBs at 10M+ vectors
- No distributed vector search without manual sharding
