Reorder PDF pages
Drag pages into a new order, remove the ones you do not want, and rotate any of them. Everything happens in your browser with the pages in front of you.
You can see the pages, so drag them
The reason hosted tools make you type page numbers is that they do not have your document — you do. Once the file is open in this page, every page is a thumbnail and reordering is dragging.
This is the operation where running locally is most obviously better, and it is why it is worth building. The alternative involves uploading a file, waiting, describing the order you want in a text field, and waiting again.
A quiet detail about how this is done
There is no function in the underlying library for moving a page — no movePage, nothing equivalent. Reordering means constructing a new document and copying the pages into it in the order you chose.
That sounds like a workaround and it is actually the better outcome. Because the document is rebuilt from the pages you kept, anything you removed is genuinely not copied across, rather than being unlinked and left behind. Removal here is real removal.
Rotation is metadata, not pixels
Rotating a page sets a value on the page object; it does not re-render anything. The result is lossless and instant no matter how large the page is, and it is stored in ninety-degree steps because that is all the format allows.
If a scan comes in upside down, this fixes it without touching the image data at all.
What the page tree actually is
The document catalog points at a node whose Kids array lists its children and whose Count records how many pages sit beneath it. Those children can be pages or further nodes, so the structure is a tree rather than a list — large documents are usually balanced into several levels so a reader can jump to page 400 without walking through 399 others.
A page knows its own position only through its place in that array. Nothing on the page records a page number; what you see printed in the corner is drawn text that someone put there.
Some properties are inherited down the tree — resources, the media box, default rotation — so a page can be missing a value that its parent supplies. This is why a page occasionally behaves differently once moved into a document with a different tree, and it is the one genuine hazard in reordering.
Answers
Can I remove pages here too?
Yes, and because the document is rebuilt from the pages you kept, removal is genuine — the removed content is not copied into the new file.
Does rotating lose quality?
No. Rotation is a property of the page, not a transformation of the image, so nothing is re-encoded.
Is there a page limit?
No. Very large documents are limited only by your own machine.