AI Features
Opensolr Search includes two AI-powered features that use your own content to generate intelligent answers. Both require a vector-enabled index (the __dense suffix, which is the default when the module creates your index).
AI Hints
When enabled, a purple "AI Hints" button appears above search results. When a visitor clicks it:
- The module collects the search query and snippets from the top results
- This context is sent to the Opensolr AI service
- An AI-generated answer streams in real time into a panel above the results
The answer is generated only from your own content — it doesn't make things up. Think of it like a smart summary that reads your top results and gives the visitor a quick answer.
The answer appears in a purple-bordered panel with text streaming in word by word. It supports Markdown formatting — headings, lists, bold, code blocks, and links all render beautifully.
Enabling AI Hints
- Go to the Search Display tab (
/admin/config/search/opensolr/display) - Check the "AI Hints" checkbox under Search Features
- Click Save
The button only appears when there's a search query and at least one result. It doesn't show on empty pages or zero-result searches.
AI Reader
AI Reader lets visitors get a quick AI summary of any individual search result. A "Read" button appears on each result card. Clicking it:
- Opens a full-screen modal
- Fetches the full document from the search index
- Streams an AI-generated summary of that specific page
This is especially useful for long documents, PDFs, and technical content where visitors want a quick overview before clicking through.
Requirements
- Your index must be vector-enabled (name ends in
__dense) — this is the default when the module creates your index - AI features use your Opensolr plan's AI quota
- Free plans include AI requests — if you hit the limit, visitors see a friendly "upgrade" message instead of an error
AI answers are generated using only your indexed content. No visitor data, search queries, or personal information is stored or used for AI training. The AI service runs on Opensolr's infrastructure.
Caching
AI responses are cached in Memcached after the first generation. If another visitor asks the same question, they get the cached answer instantly without using additional AI quota. Cache persists until your server's Memcached is restarted.