Vector database

A database or engine optimized for similarity search over embedding vectors, typically with metadata filters and hybrid lexical+vector queries for production RAG.

Vector databases store high-dimensional embeddings (e.g. 1536-d text vectors) and use approximate nearest-neighbor indexes (HNSW, IVF, etc.) for low-latency search. Operators care about hybrid search, multi-tenant isolation, replication, backup, and cost curves at billions of vectors. Choices range from fully managed SaaS to self-hosted engines and Postgres pgvector; benchmark recall@k on your co...