@twynn That tool only extracts raster/bitmap images in the PDF. It's also possible for a PDF page to include vector graphics, not embedded as image objects per se, but simply part of the stream of graphics commands that make up the page, intermixed with (and possibly including) text. Recognizing the difference between those vector graphics and plain text probably requires some kind of layout detection heuristic or machine learning model.

Hmm. Looks like someone created pdf2svgsnipper.com, as per:
reddit.com/r/pdf/comments/16ip…