PostgreSQL (pgvector)

An open source PostgreSQL extension enabling vector similarity search and storage directly alongside relational data.

This extension turns PostgreSQL into a highly capable vector database by adding a native vector data type and support for exact and approximate nearest neighbor search (using HNSW and IVFFlat indexes). It handles single-precision, half-precision, binary, and sparse vectors, allowing developers to query embeddings using distance metrics like L2, cosine, and inner product. By keeping vector search inside Postgres, engineering teams bypass the operational overhead of a separate vector database, retaining full ACID compliance, point-in-time recovery, and standard SQL JOINs.

https://github.com/pgvector/pgvector