Google converts every PDF it crawls into an HTML page before indexing, and that conversion is why PDF links pass equity, PDF text can rank in AI Mode, and image-heavy PDFs often fail to rank at all. Google's Martin Splitt and John Mueller confirmed the pipeline on the Search Off The Record podcast.

The statement came directly from Splitt, a Google Webmaster Trends Analyst, describing the indexing architecture. Google does not use a PDF viewer. It uses a converter component that sits between rendering and indexing.

How does Google actually process PDFs?

Splitt explained the mechanics in detail:

"What we are doing in Search is we have a component called a converter. And the converter component is part of what happens in indexing. It happens right after rendering and right before something potentially makes it into the index."

The converter handles more than PDFs. Splitt named rich text files, document files, Excel sheets and Lotus Notes documents as formats that pass through the same conversion layer. The output is HTML.

"So we are taking that and we are producing an HTML version of it so that the processing in the rest of the pipeline pretty much only works with HTML."

That's a confirmed Google statement, not tool-signal speculation. The entire downstream ranking pipeline — the part that evaluates content, follows links and builds the index — sees PDFs as HTML pages.

Do links inside PDF files work?

Yes. Mueller confirmed that links embedded in PDFs function like normal HTML links, precisely because the HTML conversion preserves them.

"So basically when Google sees a PDF file, it converts it into something that it has more practice with, kind of a web page. And I guess that's also why things like links also work in PDF files."

Splitt confirmed it:

"Yep. Because we convert it into an HTML page and then the rest of our systems are like, 'I know how to deal with an HTML page,' and they just do their stuff."

He added that the conversion stays "as true as is possible" to the original format, so links survive the transformation. For SEO teams, this is the clearest Google confirmation available that PDF link building is not wasted effort.

What happens to images inside PDFs?

Here the picture is murkier. Splitt expressed genuine uncertainty — this is Google's honest guess, not a documented behavior.

Images in PDFs lack URLs, and the normal image pipeline works from URLs. Splitt said Google likely detects the image exists but probably cannot process it for image search the way it processes a standard web image.

"I think we understand there is an image, but I don't think we can actually process it like a normal image for image search. But that might change or might have changed; that's my current understanding of how that pipeline works."

Practical takeaway: do not count on PDF-embedded images appearing in Google Images. He flagged that this could change, so it is worth monitoring.

What are the size limits?

Mueller asked directly about file size limits. Splitt's answer was partial. Fetching a PDF likely carries a larger limit than HTML, similar to how Google treats videos and images. But he did not know the exact number.

"I assume that we have a larger file limit, just the same as we do for videos and images. But I don't know the specific limit for PDF files."

Once the PDF converts to HTML, normal HTML limits apply, according to Splitt. Those HTML limits, discussed in earlier podcast episodes and Google documentation, originate in the crawling component of the pipeline. Splitt also said the current PDF limit could rise if Google detects publishers uploading larger files.

Can PDFs rank in AI Mode?

Confirmed, yes. Splitt stated PDFs can rank in AI Mode for the same reason they rank in regular search — Google indexes the HTML conversion. Mueller reinforced the point:

"So they can also appear in AI overviews, AI mode, all of those things."

One caveat matters for anyone publishing scanned documents. Some PDFs are essentially images. If you cannot highlight and copy the text into a text editor, Google likely cannot convert it into indexable HTML. That copy-paste test is the fastest diagnostic available for whether a PDF is rankable.

Should publishers still use PDFs?

Splitt endorsed PDFs where the format earns its place: print-specific formatting, interactive forms, registration documents. His example was a holiday lodge publishing a registration form, and restaurants uploading print-ready menus.

"If you have a PDF for whatever reason, go for it. Restaurants oftentimes have the printers create the menus that they actually physically print on paper as a PDF as well, and you can just upload that to your website. It's better than not having your menu on the website at all."

Is HTML still better for SEO?

Mueller said yes, on pragmatic grounds. He prefers HTML pages over PDFs when both options exist, because HTML is easier for search systems to process and easier for users to interact with.

"If the option is have a menu as a PDF or don't have a menu at all, then yes, I would prefer to have a menu. But if the option is menu as a PDF versus a normal web page, I kind of like normal web pages."

His bottom line for site owners: if content works equally well in HTML, publish it in HTML. That advice echoes the founding principle of SEO — reduce friction for crawlers and indexers wherever possible.

For site owners with large PDF libraries — spec sheets, white papers, menus, manuals — the next signals to watch are whether Google starts indexing PDF-embedded images and whether it raises the PDF size limit as larger documents proliferate.