We don't. We extract the content and present it as a single document.
Page anchors can be used for navigating between sections. We present a table of contents that makes this easy. For instance:
https://papertohtml.org/paper?id=6f9fc51102cf49bff4f4e2b3367...