How Google Search treats PDFs
Google can index PDFs, rank them in regular search results, and sometimes reference them in AI Overviews or AI Mode. In an October 2026 Search Off the Record episode, Google’s Martin Splitt and John Mueller explained how Search handles the format.
What Google can read
A PDF is a container rather than a plain-text web page. Google converts it into an HTML representation so its indexing systems can process the content. That includes text links. Google says it may follow links inside a PDF and use their ranking signals.
Images need more thought. Graphics embedded in a PDF do not have their own URLs in the document. If an image matters on its own, publish it as an image on an HTML page too.
Google also says a PDF’s title metadata and the text of links pointing to the file can influence its title in search results. Give the file a descriptive title before uploading it, and link to it with words that describe the document.
Make the text readable
A scanned PDF may contain pictures of words rather than selectable text. Google may extract those words with OCR, but a proper text layer gives it a more reliable starting point. Try selecting a sentence and pasting it into a text editor. That is a useful check, though it cannot prove how Google will index the file. Password-protected or encrypted PDFs generally cannot be indexed.
File size and AI results
Googlebot’s current documentation says it fetches the first 64 MB of a PDF, compared with the first 2 MB of other supported file types, including HTML. Content beyond those fetch limits is not sent on for indexing consideration. That larger allowance is a crawl limit. Do not make a PDF bigger just to use it.
PDFs can appear in Google’s AI search features because Google can process and index them. The podcast did not establish a special AI ranking benefit for PDFs or a reason to convert content to PDF for “GEO”. It also did not establish how other AI companies use PDF content.
Duplicate versions and updates
If you publish the same material as both HTML and PDF, Google may choose which version to show. Decide which you prefer. Google’s canonicalization guidance explains how to signal an HTML page as the preferred version with a rel="canonical" HTTP response header on the PDF. If you want the PDF kept out of Search entirely, serve it with an X-Robots-Tag: noindex header.
Google tends to revisit PDFs less often than HTML pages. Replacing a PDF at the same URL may therefore leave an older version in search results for a while. This is also worth considering if your CMS automatically creates a printable PDF copy of every page.
Choose the format for the reader
PDFs are useful for printable forms, reports, manuals, and documents whose fixed layout matters. HTML usually works better for articles, restaurant menus, prices, and other information people read on phones or update often. A PDF download can sit alongside the page when readers need one.