Why a split PDF sometimes “bloats”
People expect page 1 of a 40 MB file to be about 1 MB. That is only true if every page carries its own images. In a real PDF, a 4 MB JPEG of a letterhead can live once in the resources and be referenced by all 40 pages. Extract page 1 and that JPEG still has to come along, because the page paints it. The one-page file might be 4.2 MB. That is not a bug and it is not quality loss — it is how resource sharing works.
We copy page dictionaries; we do not resample. Text stays selectable. A 300 ppi scan stays 300 ppi. If you want the extract smaller, run Compress after Split.
Two download modes
One PDF per selected page is for portals that want “page 2 only” as its own file — you get a ZIP when there are several. One PDF of the selected pages is for “send 3–7 as a packet.” Thumbnails render with PDF.js in this tab; they are previews, not the bytes we write out.