AI Rag Search Summary

Introduction

AI RAG Search Summary turns your Search API vector search into an intelligent Q&A system. Users ask questions in plain language and get an AI-generated summary answer along with the actual source documents used to verify information and explore deeper. No more scrolling through search results.

Choose how sources appear: use any Drupal view mode (teaser, full, custom) to render source nodes with your existing templates, or use the default JSON format. Style with CSS variables to match your theme.

Post-Installation

Prerequisites

• Have Drupal Search API configured with AI Search backend
  • New to AI Search? Read the docs found here AI Search project page for setup instructions
  • Video tutorials:
    • AI Search Setup Guide
    • Configuring Vector Search
• Content already indexed with embeddings

2. Create Configuration

1. Go to /admin/config/ai/rag-search-summary
2. Click "Add configuration"
3. Select your Search API index
4. Choose AI provider: openai__gpt-4o
5. Customize prompt, temperature, caching, and rate limits
6. Save

3. Add Block

1. Place the "RAG Search Summary" block via /admin/structure/block
2. Select your configuration
3. Customize UI text and display options

That's it! The block is now ready to use.

Customization

Styling with CSS Variables

The React component uses CSS variables for easy theming without touching component code per block config:

--rag-color-primary: #0a3f5a;
--rag-color-primary-hover: #0076a9;
--rag-color-primary-dark:  #0a3f5a;
--rag-color-text-primary: #212529;
--rag-sources-columns-mobile: 1;
--rag-sources-columns-tablet: 2;
--rag-sources-columns-desktop: 2;
--rag-font-heading: "Forma DJR UI", Arial, sans-serif;
--rag-text-base: 18px;