Why is my PDF upside down?
You open a PDF and the pages are upside down, or turned on their side. You rotate it in the viewer, it looks correct, you close it — and next time it is wrong again. Or you send it to somebody and they tell you it arrived sideways, even though it looks perfectly fine on your screen.
Both of those are the same underlying quirk, and it is worth two minutes to understand because it decides which fix actually works.
A PDF page stores its rotation separately from its content
Inside a PDF, each page carries a value called /Rotate. It is a single number — 0, 90, 180 or 270 — and it means "turn this page by this many degrees before displaying it". The page content itself is not moved. The instruction sits alongside it.
So a page can be upside down in two completely different ways:
- The content is the right way up and /Rotate says 180. The file is fine; it is being displayed rotated because it was told to.
- The content was scanned or generated upside down and /Rotate says 0.The instruction is correct and the pixels are wrong.
They look identical on screen. They behave completely differently when you print, extract text, or open the file somewhere else.
Why rotating in a viewer does not stick
This is the part that frustrates people most. In many PDF readers — including the preview panes built into browsers, Windows and macOS — the rotate button is aview control. It turns the page on your screen and changes nothing in the file. Close the document and the setting is gone.
Full editors like Acrobat do modify the file, but only if you rotate and then save. Rotating and closing without saving discards it, which is easy to do when the page already looks right.
The test: rotate the page, close the file completely, and reopen it. If it is wrong again, you changed the view rather than the document.
Why it looks right for you and wrong for them
Most software respects /Rotate. Some does not. Older printer drivers, certain corporate document systems, some text-extraction tools and a number of phone apps either ignore the value or apply it inconsistently.
So a file with content stored upside down and /Rotate set to 180 displays correctly in every reader that honours the flag, and upside down in every one that does not. You see a perfectly normal document. The recipient sees an upside-down one. Neither of you is doing anything wrong.
Printing is where this bites hardest, because printer drivers are exactly the kind of older software most likely to ignore the flag.
Where the rotation comes from in the first place
Scanners are the most common source. A document fed in the wrong way round, or a duplex scanner where the reverse side comes through inverted, produces alternating pages — odd pages upright, even pages upside down. That specific pattern almost always means duplex scanning.
Phone scanning apps guess orientation from the photo, and guess badly on documents with little text, forms, or anything photographed at an angle.
Landscape content in a portrait page. Spreadsheets and slides are often exported as a portrait page with the content rotated 90 degrees, so the file "is" portrait but reads sideways. This is deliberate rather than a mistake, and it is why so many exported reports open on their side.
Faxes and older document systems frequently produce fixed-orientation output with no rotation flag at all.
How to fix it permanently
You need a tool that writes the change back into the file rather than adjusting the view. Our Rotate PDF tool does that — it modifies the page's rotation and gives you a new file with the change saved in. Nothing is uploaded; the edit happens in your browser.
A few things worth knowing when you do it:
- Rotate every page, or only some? If the whole document is wrong, rotate everything. If only alternate pages are wrong, that is duplex scanning, and you want to turn just the affected ones by 180 degrees.
- Check the direction. Sideways pages need 90 degrees, but which way is not obvious from a thumbnail. If the result is upside down, you went the wrong way — rotate 180 more.
- Reopen the saved file to confirm. Same test as before. A fix that survives closing and reopening is a real fix.
When rotating is not enough
If a scan is skewed — a few degrees off, rather than a clean 90 or 180 — no rotation tool will straighten it. PDF page rotation only handles right angles. A slightly crooked scan has to be rescanned, or straightened as an image before being turned back into a PDF.
For a crooked phone photo of a document, the practical route is to fix the image first and rebuild the PDF: crop androtate the photo, thenconvert it back to PDF. It is more steps, but it is the only thing that addresses the actual problem.
A note on text and searchability
Rotating a page does not affect whether its text can be selected or searched. That depends on whether the PDF has a text layer at all — a scan is a picture of words, and stays a picture of words however you turn it.
If you need the words out of a scanned document, orientation is a separate problem from extraction. Our PDF to Text tool reads an existing text layer, and will tell you plainly when a file does not have one.
The short version
Rotation is an instruction stored in the file, not a property of the image.
Viewer rotate buttons usually change the view, not the document.
Test any fix by closing and reopening the file.
Alternating upside-down pages means duplex scanning.
A skewed scan is a different problem and needs a different fix.