I solved my problem. It's quick and dirty, but it works.

I used the URL in the last post as my base URL, set depth to 2 and included
webp and pdf in the filter.

It downloaded ~40MB of data, which is a lot more than those pages, but it
works and for now, it's enough.