Pgvector is an open-source tool designed for vector similarity search within Postgres databases. With Pgvector, you can conveniently store your vectors alongside other data, offering support for both exact and approximate nearest neighbor search. The tool handles various distance metrics, including L2 distance, inner product, and cosine distance. It seamlessly integrates with any programming language that has a Postgres client.
Notably, Pgvector ensures ACID compliance, point-in-time recovery, JOINs, and all the excellent features inherent to Postgres. This tool simplifies the process of incorporating vector similarity search functionality into your database, making it efficient and accessible for users working with diverse data types and programming languages.