It extracts the metadata, but not the body, yes. Retrieved websites are generally frowned upon in my field (medicine, virology), and that's why I never really needed this. We're pandoc-ing our internal MD files to PDF and attach them to the relevant HTML pages, which makes this somewhat a crutch but circumvents the "9 out of ten reviewers hate websites" problem.

I print the webpage to a PDF and store it that way, then fix up the metadata as I see fit. And that is a good feature - make a request on the Github repo and see if the lead developer responds.