Search inside the files

With the text extracted, searching your files is just searching. The difference a visitor feels: the query can match words that appear only inside a document, not just its title.

A phrase from deep in a paper

Here's the finished example. "Scaled dot-product attention" appears nowhere in any title or summary — it's a phrase from inside one specific paper. Search it and that paper surfaces first, with the matching passage pulled straight out of the PDF so the reader can see why it matched:

The document search demo: searching 'scaled dot-product attention' surfaces the Attention Is All You Need paper, showing the matching passage from inside the PDF

The call your page makes

It's the same search endpoint every other lesson uses — the extracted text is simply part of what gets searched, and each hit comes back with the record's fields ready to display:

GET https://api.searchstack.dev/search/Demo/docs/1?query=scaled dot-product attention
X-API-Key: {your key}

That's the lesson. Files into a media store, a record pointing at each one, and the text inside them extracted and indexed for you. Whatever your PDFs are — manuals, contracts, papers, reports — your visitors can find the right file from a phrase buried on page five.

Go deeper: Searching — queries, filters, and what comes back with a hit in the reference.