pgvector + Postgres review and pricing

pgvector is the open-source Postgres extension that adds vector types, distance operators, and approximate-nearest-neighbor indexes (IVF and HNSW) to any Postgres instance. The honest framing: for most production RAG workloads under 5-10M vectors, pgvector on a well-tuned Postgres with HNSW indexing is the most cost-effective answer, particularly because most buyers already pay for a managed Postgres (RDS, Aurora, Cloud SQL, Azure Database for PostgreSQL, Supabase, Neon, Crunchbridge). The decision to add a dedicated vector DB should be driven by scale, latency at very high concurrency, or multi-tenancy requirements, not by vendor pitch. Trade-offs: at very large scale (tens to hundreds of millions of vectors) dedicated DBs still win on latency and operational ergonomics.