What rotation actually changes in the file
A PDF page carries a /Rotate entry — 0, 90, 180, or 270 — that tells every viewer how to orient the page for display. Rotating a PDF here just updates that number: the page’s content streams, fonts, and images are byte-for-byte untouched. That’s why the operation is instant even on huge files, why quality can’t degrade, and why the file size barely moves. It’s the same mechanism your scanner’s software would have used if it had guessed the orientation correctly in the first place.
The flip side of this design explains a common confusion: the rotate button inside a PDF viewer usually changes only the current viewing session, not the flag in the file. The document looks fixed on your screen, then arrives sideways in someone’s inbox. A tool that rewrites the /Rotate entry — like this one — is what makes the fix stick.
Mixed-orientation documents
Real-world documents often legitimately mix orientations: a portrait report with a landscape financial table, or an appendix of wide charts. When normalizing a file like that, don’t blanket-rotate everything — rotate only the pages that are actually wrong, using the per-page thumbnails. Viewers, printers, and OCR all handle mixed orientation fine as long as each page’s flag matches its content. And if the pages also need reordering or pruning while you’re at it, the PDF Page Organizer combines rotation with drag-and-drop restructuring in one pass.