Ai Recipe Image Search Vector

Takes the AI Image Classification recipe and makes its results searchable by building a Search API index, backed by your preferred vector database of choice.

What it does

The AI Image Classification recipe enriches image media with AI-generated metadata - an image description, alt text, and taxonomy tags. On its own, that metadata sits on the media entities but is not easy to search.

This recipe adds the missing piece: it applies the image classification recipe and then creates a Search API server and index that expose those AI-generated fields (plus the image's own properties) as a searchable index. Indexing happens automatically as content is created and updated.

Designed to pair with an agent. This recipe is primarily a building block for an AI Agent recipe that depends on it - the agent searches this index to find relevant images. That said, the index is a standard Search API index, so it is fully usable on its own (for Views, custom search pages, or any other Search API consumer) without any agent.

What gets installed

• The AI Image Classification recipe (applied as a dependency).
• search_api and ai_search — Search API and its vector backend of choice.
• automated_cron — so content still gets indexed even when no system cron job is configured (indexing runs during cron).

Using it on its own

1. Apply the recipe.
2. Make sure your images have been classified (the classification recipe generates descriptions and tags).
3. Let cron run, or index manually with drush search-api:index image_database_index_vector.
4. Build a View on the Image Database Index (Vector), or query the index from your own code, to search and filter images by their AI description, tags, alt text, and more.

Requirements

• drupal/ai_recipe_image_classification ^1.0
• drupal/search_api ^1.39
• drupal/ai_search ^1.3-alpha1