Qdrant Vector Database
AdoptFactorial rates Qdrant as Adopt (2026). Qdrant is our vector database for Drupal projects.
Qdrant is an open-source vector database written in Rust. It stores embeddings together with a JSON payload, and searches them by cosine, dot product or Euclidean similarity, filtered by payload fields. It runs as a single Docker container, as a cluster or as Qdrant Cloud.
It is a common choice for semantic search and retrieval-augmented generation when vectors should live in a dedicated service instead of the application database, and it integrates with most AI frameworks, including Drupal's AI module through a vector database provider.
Updates
Qdrant is our vector database for Drupal projects. We maintain the Qdrant provider for Drupal's AI module, which plugs Qdrant into AI Search and other vector-aware modules; it supports Drupal 10.2 and 11 and is used on more than 100 sites. Our controlling app embeds AI-generated summaries of projects and epics into Qdrant to power its semantic search, "similar epics" and the search_summaries tool of its MCP server. The summaries themselves are generated by FlowDrop workflows. Semantic results are merged with plain lexical matches, because exact identifiers such as ticket keys are not something embeddings find reliably.
Where an application already runs on PostgreSQL and has no Drupal integration to reuse, we use pgvector instead.