About

Provides a Weaviate vector database provider for Drupal AI Search.
Behind the scenes, it uses the Weaviate PHP client.
Works with Drupal core 10.3 or 11.x+.

Weaviate

Weaviate is an open source vector database (VDB) that stores both objects and vectors, allowing for the combination of vector search with structured filtering with the fault tolerance and scalability of a cloud-native database​. See more about Weaviate on https://weaviate.io and https://github.com/weaviate/weaviate.

Features

  • Weaviate connectivity and health checks.
  • API-key authentication through the Drupal Key module.
  • Configurable request timeout and additional HTTP headers.
  • Collection lifecycle operations using Weaviate schema classes.
  • Database-scoped Weaviate class names.
  • Configurable cosine, Euclidean, and inner-product similarity metrics.
  • Batch vector indexing with deterministic object IDs.
  • Server-side deletion by Drupal entity ID, including chunked embeddings.
  • Vector and non-vector GraphQL searches.
  • Drupal AI Search server and index isolation filters.
  • Dynamic metadata properties for indexed Search API fields.

This project is not affiliated or related to Weaviate company.

Supporting organizations: 
Initial development

Project information

Releases