pgvector

pgvector is the open-source PostgreSQL extension that enables efficient storage and vector similarity search of high-dimensional embeddings.

pgvector transforms PostgreSQL into a powerful vector database, allowing you to store and query machine learning embeddings directly alongside your relational data . It introduces the `vector` data type and supports critical operations like k-nearest neighbor (k-NN) search using standard SQL . For performance, it offers approximate nearest neighbor indexing with HNSW and IVFFlat algorithms . You can leverage multiple distance metrics (L2 distance, inner product, cosine distance) to power modern AI applications like semantic search and recommendation systems, all within the robust Postgres environment .

https://github.com/pgvector/pgvector