Rotate PDF

The difference between turning a page and turning your head

Every PDF reader has a rotate button, and in most of them it changes only how the document is displayed on your screen. Close the file, send it to somebody else, and it is sideways again. What you want is for the rotation to be part of the document.

A PDF page carries a /Rotate entry — a number of degrees, always a multiple of 90 — that says how the page should be turned before it is shown. This tool sets that entry. Everything that opens the file afterwards, including a printer, sees the page the way you left it.

Nothing is redrawn, so nothing is lost

The page content is untouched. No image is re-encoded, no text is re-rendered, no resolution is lost, and the file barely changes size. Rotating a hundred-page document takes about as long as saving it, because that is essentially all that happens.

Rotation is cumulative. A page that was already turned 90° and is turned 90° again ends up at 180°, which is what you would expect and is not always what other tools do.

Turning only some of the pages

Leave the pages box empty and every page turns. Otherwise use the same syntax as everywhere else on this site: 1-3, 7, 12-. When you load a file, the line above the controls tells you how many pages are landscape and how many are portrait, which is usually enough to work out which ones came in wrong.

The common case is a scan where the feeder took a few sheets in the wrong orientation. Turn those and leave the rest alone.

What it does not fix

A page that is skewed rather than rotated. If a sheet went through the scanner at four degrees, this cannot help — rotation in a PDF is only ever a multiple of 90. That needs a deskew, which means re-rendering the page, which means losing the text. Scanner software is the right place for it.

Mirrored pages. A page scanned through the back of the sheet is flipped, not turned, and no amount of rotation will make it readable.

Signed documents. Rotating a signed PDF invalidates the signature, as any modification would.

Why scanned pages come in the wrong way round so often

Three separate causes, and knowing which one you have saves time.

The sheet went in sideways. A document feeder takes whatever orientation the stack was in, and a mixed stack produces a mixed document. This is the case this tool fixes: turn the affected pages and save.

The scanner auto-rotated based on the text and got it wrong. Most scanning software tries to detect which way up a page is by looking at the writing on it. On a page that is mostly a table, a photograph, or a language it does not recognise, it guesses, and it guesses wrong about as often as you would expect. Again, fixable here.

The page is landscape and was always meant to be. A wide table or a spreadsheet export is landscape on purpose. Before turning everything, check whether the "sideways" pages are actually correct and it is your reader that is displaying them oddly.

The line above the controls counts landscape and portrait pages when you load a file, which usually answers the question immediately: a document with 38 portrait pages and 2 landscape ones almost certainly has two pages that went in wrong.

Doing this to a lot of documents

This page handles one document at a time. If you have a folder of scans all rotated the same way, the tool will still be faster than opening each in a PDF editor, but it is a page-at-a-time workflow and it is honest to say so.

For genuinely bulk work, the command-line tools qpdf and pdftk both rotate pages and can be run over a whole directory in one line. They are free, they run on your own machine, and for a hundred files they are the right answer. This tool exists for the far more common case of one document, right now, on a machine where you cannot install anything.