Can't say much about this, the docs seem to talk about form fields and mention JS, so it might not be what the OP wanted.

Anyway the OP was talking about "replacing placeholders".

With a PDF version where allowed fields are completely filled with blind text in a dedicated font, pdftohtml -xml could easily be used to parse and grep the absolute positions for these text boundaries.

And plotting text by absolute positions is trivial with PDF::Reuse. (of course into a PDF without blind text)

But as I already said, one has to care about line breaks and staying within bounding boxes, cause PDF is not HTML, it's a print format!

see also: Parsing PDFs by text position?