Announcing pgvector 0.7.0 support in Aurora PostgreSQL

The article announces the release of pgvector 0.7.0 in Amazon Aurora PostgreSQL-Compatible Edition, an open-source extension for PostgreSQL that enables storing vector embeddings in the database and provides vector similarity search capabilities for semantic search and retrieval-augmented generation (RAG) in generative AI applications.

Specifically, the article covers:

• New features in pgvector 0.7.0, including parallelism for faster index building, new vector data types (halfvec and sparsevec), and support for indexing binary vectors
• Benefits of the new features, such as reduced index storage size, lower index build time, and increased maximum dimensions for vectors
• Availability of pgvector 0.7.0 in Aurora PostgreSQL clusters running specific versions in all applicable AWS Regions except China regions
• Information on how to initiate a minor version upgrade to access pgvector 0.7.0
• Overview of Amazon Aurora's high performance, availability, and compatibility with PostgreSQL