Scolta
Scolta is a Drupal search backend that uses Pagefind to provide fast, AI-enhanced search directly in the visitor's browser. It offers configurable relevance scoring and optional AI-powered features like query expansion and result summarization, all without needing external search servers.
Scolta is a Search API backend that delivers fast, AI-enhanced search for Drupal sites. Built on Pagefind, it adds configurable relevance scoring, AI-powered query expansion, and result summarization — all without requiring external search infrastructure like Solr or Elasticsearch. Search queries resolve entirely in the visitor's browser against a pre-built static index.
Features
Scolta provides a complete search solution as a Search API backend:
- No external search server required. Pagefind builds a static index at publish time and runs a WASM search engine in the browser. No Solr, no Elasticsearch, no managed service to configure or pay for.
- Configurable relevance scoring. Re-ranks Pagefind results with tunable boosts: title match weight, content match weight, recency decay curves, phrase-proximity multipliers, and exact title match boosting.
- AI-powered query expansion (optional). When configured with an LLM provider, Scolta rewrites queries for better recall, generates result summaries, and suggests follow-up questions. Supports multiple backends including the Drupal AI module (48+ providers), Anthropic, OpenAI, and OpenAI-compatible endpoints. Base search works without any AI provider.
- Client-side search performance. Pagefind's Rust/WASM engine handles indexing and search. Queries run in the visitor's browser with no server-side search overhead.
- Full Search API integration. Works as a standard Search API backend. Compatible with Views and other modules that consume Search API indexes.
- Drush commands for index management. Build, rebuild, export, and monitor your search index from the command line. Supports chunked builds, resumable indexing, and memory budgets for large sites and shared hosting.
Post-Installation
After installing via Composer (composer require tag1/scolta-drupal), enable the module and configure it at Administration → Configuration → Search and metadata → Search API. Create a new server, select Scolta Pagefind as the backend, and create or update an index to use it. Then build your search index:
drush scolta:buildPlace the Scolta Search block on your site via Structure → Block Layout. To enable AI features (query expansion and summarization), configure an LLM provider in the Scolta settings form — this step is optional.
Additional Requirements
- Drupal 10.3+ or Drupal 11
- PHP 8.1+
- Search API module
- The Scolta PHP library (
tag1/scolta-php) is installed automatically as a Composer dependency and includes the Pagefind WASM runtime and frontend assets.
Recommended modules
- Drupal AI — Multi-provider AI integration with Key module support, rate limiting, and token tracking. When selected as the AI provider in Scolta's settings, routes query expansion and summarization through the Drupal AI module's plugin system instead of Scolta's built-in AI client.
Similar projects
- Search API Solr — Requires running and maintaining an Apache Solr server. Proven at very large scale. Scolta trades that server-side power for zero infrastructure overhead and client-side search.
- Search API Elasticsearch — Requires an Elasticsearch cluster. Same trade-off as Solr: more operational overhead and more features at scale.
- Database Search (search_api_db) — Ships with Search API, no external dependencies. Scolta adds AI-powered query expansion, configurable relevance tuning, and client-side WASM search that database search cannot provide.
Supporting this Module
Scolta is developed and maintained by Tag1 Consulting.