Skip to main content

AI and vector search in Aiven for PostgreSQL®

Use the pgvector extension to power AI-based vector search in your Aiven for PostgreSQL® service.

Why use pgvector instead of a separate vector database​

Because pgvector runs inside PostgreSQL rather than as a standalone system, your embeddings share the same transactions, backups, and access controls as the rest of your data. Querying relational data and vector similarity together also means you skip building and operating a pipeline to keep a separate vector database in sync with your source of truth.

Things to know​

  • pgvector is available as an extension on every PostgreSQL major version Aiven supports, so you can adopt it without planning a version upgrade first.
  • For larger vector workloads, the pgvectorscale extension complements pgvector with additional indexing. See Extensions on Aiven for PostgreSQL® for supported versions.

Related pages