pgvector is an open-source extension for Postgres that allows you to store, query, and index vector embeddings directly within your database. It is commonly used for semantic search, retrieval-augmented generation (RAG), and AI/ML applications that require similarity search over high-dimensional data.
pgvector is supported on EDB Postgres Advanced Server (EPAS), EDB Postgres Extended Server (PGE), community PostgreSQL, and EDB Postgres AI for WarehousePG. For supported versions and platform requirements for EPAS, PGE, and PostgreSQL, see the Compatibility page. For WarehousePG, see the pgvector in WarehousePG documentation.
This section covers installing and configuring pgvector, working with vector data and indexes, integrating with application drivers and AI/ML frameworks, securing vector workloads, and day-to-day maintenance.