New AI search and vector database tools released for Drupal and Postgres

New technical tools have been introduced to facilitate AI-driven search and Retrieval-Augmented Generation (RAG) capabilities.

Varbase has released an experimental Drupal recipe, drupal/varbase_ai_search_base, designed to create a vector index of a site's content. This tool serves as the retrieval component for the ‘Ask AI’ feature, allowing sites to index nodes and pages through a default vector database provider. Supported providers include MariaDB, Milvus, and SQLite.

Additionally, pgvector has been highlighted as an open-source extension for PostgreSQL. It enables vector similarity search directly within existing Postgres databases, allowing users to manage embeddings, document chunks, and user permissions within a single SQL query. This integration supports standard database features such as transactions, backups, and point-in-time recovery without the need for a separate database synchronization.