Tag1 Guide Explains Scolta AI Search Setup for Drupal

Closing the gap between what users type and what Drupal content actually says is the focus of Jeremy Andrews’s Tag1 guide, How to Scolta on Drupal. The post presents Scolta as an open-source Drupal module that improves search by expanding user queries and connecting conceptual searches to relevant content that may not contain the same keywords.

The tutorial demonstrates Scolta using The Athenaeum, a Drupal 11 demo site built from more than 6,000 English Wikipedia Featured Articles. Andrews explains that Scolta integrates with Drupal via the Search API module and uses Pagefind to build a static, browser-based search index, thereby avoiding the need for Solr, Elasticsearch, or a separate search daemon. The setup path covers installing Scolta with composer require drupal/scolta, enabling it with drush en scolta, configuring a Search API server, building the index with drush scolta:build --force, and connecting an AI provider through amazee.ai, a direct provider key, or the Drupal AI module.

The post also explains how Scolta uses site-type presets, scoring controls, metadata, and sortable fields to tune results. For the encyclopedia demo, the “Recipe & Content Catalog” preset disables recency, increases full-title matching, raises body-text weighting, and broadens the result set before re-ranking. Andrews identifies the site description as the most important quality lever because it gives the AI model context for query expansion, and shows how indexed metadata such as word count, reference count, date, taxonomy, and custom sortable fields can guide both search ordering and AI-generated overviews.