Reordering, rotating and splitting pages without touching their content
Most of what people call "editing a PDF" is not editing at all — it is rearranging. Dropping the blank page the scanner added, turning the sheet that came in sideways, pulling three pages out of a forty-page contract, putting the annex before the appendix. None of that touches a single letter of the content, and that is precisely why it can be done safely and quickly, right here.
Thumbnails first, decisions second
Tools that ask you to type page ranges into a box invite mistakes: you delete 12–14 and discover later that the numbering printed on the page was offset from the numbering in the file. Here the document is rendered page by page into thumbnails before you touch anything, so you act on what you can see. The rendering uses PDF.js — the same engine Firefox uses to display PDFs — running locally.
Nothing you do in the grid changes the file. Reordering, rotating and removing all happen to an in-memory list; the document is only rebuilt when you press one of the save buttons. Reset puts everything back, and your original file on disk is never modified either way.
Rotation is a property, not a redraw
A PDF page carries a rotation value that viewers obey when displaying it. Rotating here changes that value, which means the page is not re-rendered and loses nothing — and, crucially, the fix travels with the file. That is the difference between rotating in a viewer, where the correction usually lasts only as long as that window stays open, and rotating in the document, where whoever opens it next sees it the right way up.
Two ways to save
One PDF keeps whatever pages survived, in the order you arranged them. That covers deleting, reordering, rotating, and extracting a subset — remove every page you do not want and save what remains.
One file per page breaks the document apart, numbering the files so alphabetical order matches the original order. Useful when each page has a different destination: separate invoices from a batch, individual certificates from a printed run, one page per form to be signed by different people.
What this does not do
It does not edit the text inside a page, and no honest browser tool does. A PDF does not store paragraphs; it stores instructions to draw glyphs at coordinates, with fonts embedded as subsets containing only the characters used. Changing a word means re-extracting the font, re-laying the line and reflowing the paragraph. Tools that claim otherwise usually rasterise the page and draw over it, which silently destroys the searchable text layer.
It also will not open password-protected documents, and does not try to. And it does not compress: pages are copied as objects, so the result weighs about as much as the pages it kept. If you need several documents joined instead of one taken apart, that is the merge tool.