How to fix upside-down pages in a scanned pdf
- Step 1Identify which pages are upside down — Open the scan in any PDF viewer and note the page numbers that read upside down (text flowing bottom-to-top, header at the foot of the sheet). Write them down — the tool has no thumbnail picker, so you select pages by typing their numbers.
- Step 2Open the Rotate tool and drop the scan in — Load your file into the PDF Rotate tool. Parsing happens in your browser; the page count appears next to the filename once the PDF is read so you can sanity-check it matches your viewer.
- Step 3Choose the 180° rotation angle — Open the Rotation angle dropdown and pick
180°. That is the correct turn for an end-for-end flip. (90° clockwiseand270° (90° counter-clockwise)are for sideways pages — see the sideways-pages guide.) - Step 4Type the upside-down page numbers — In the Pages field, enter the numbers comma-separated, e.g.
3, 4, 7. Page numbers are 1-based (the first page is1). Leave the defaultallif the entire scan is flipped. Note: hyphen ranges like3-7are NOT supported here — list each page. - Step 5Process the file — Click Process. The engine adds 180° to each named page's existing rotation and re-saves the PDF. Because the rotation is additive, run it once — re-running on the same pages would turn them back over.
- Step 6Download and verify in a viewer — Download the corrected PDF and open it. Confirm the previously-flipped pages now read correctly and the others are untouched. If any page is still wrong, note its number and run a second pass with the right angle.
Rotation angle options (the three real dropdown choices)
These are the exact three options in the Rotation angle dropdown. There is no free-text angle field and no separate counter-clockwise toggle — 270° is the counter-clockwise turn.
| Dropdown label | Stored value | Use it when | For upside-down scans? |
|---|---|---|---|
90° clockwise | 90 | A page is sideways with its top edge on the left | No — this leaves the page sideways |
180° | 180 | A page is flipped end-for-end (header at the bottom, text reads bottom-to-top) | Yes — this is the correct turn |
270° (90° counter-clockwise) | 270 | A page is sideways with its top edge on the right | No — this leaves the page sideways |
What the Pages field accepts
The Pages input is split on commas and each token is parsed as an integer; tokens that are not numbers are dropped. It does NOT understand hyphen ranges.
| You type | What the tool does | Result |
|---|---|---|
all (default) | Rotates every page in the document | Whole scan turned 180° |
3, 4, 7 | Rotates pages 3, 4 and 7 (1-based) | Three named pages turned |
3-7 | Parses as the single integer 3 (the -7 is discarded by integer parsing) | Only page 3 turns — a common surprise |
3, ,abc, 7 | Empty and non-numeric tokens are dropped | Pages 3 and 7 turn |
99 on a 10-page scan | Requests page index 98, which does not exist | The job errors — keep page numbers within the document |
Free vs Pro limits for the Rotate tool
Single-source limits from the JAD Apps tier table. Rotate is a single-file tool, so the batch count rarely matters.
| Limit | Free | Pro |
|---|---|---|
| Max file size per job | 2 MB | 50 MB |
| Max pages per PDF | 50 | 500 |
| Files per job | 1 | 5 |
| Upload to server | Never (in-browser) | Never (in-browser) |
Cookbook
Real before/after states for upside-down scans, showing the angle and page list to use. The tool reports the page count of the output — the simplest sanity check.
A few flipped pages in an ADF scan
An auto-document-feeder pulled three sheets in upside down. The rest of the scan is fine, so you name only the bad pages and turn them 180°.
Before (10-page scan): pages 1,2,5,6,8,9,10 → upright pages 3,4,7 → upside down Rotation angle: 180° Pages: 3, 4, 7 After: all 10 pages upright; output still 10 pages
The entire document is upside down
A phone scanning app captured every page rotated 180° (the phone was held the wrong way up). Turn the whole file in one pass with the default page selection.
Before: every page reads bottom-to-top Rotation angle: 180° Pages: all (the default) After: every page upright
Avoiding the hyphen-range trap
You meant to rotate pages 4 through 8, but the Pages field does not parse hyphen ranges. List each page number explicitly instead.
Wrong: Pages: 4-8 → only page 4 rotates Right: Pages: 4, 5, 6, 7, 8 (If you genuinely need range syntax for a different operation, the Delete-pages and Extract-pages tools do support ranges — Rotate does not.)
A page already flagged 180° by the scanner
Some scanners write a 180° rotation flag into the page rather than re-orienting the raster. Because this tool ADDS your angle to the existing one, adding another 180° lands the page back at 0° — exactly what you want.
Page 5 current /Rotate: 180 You apply: 180° New /Rotate: (180 + 180) % 360 = 0 → upright Lesson: run the rotation ONCE. A second 180° pass would flip page 5 back upside down.
Rotate first, then re-OCR a scanned record
If you OCR a scan while pages are upside down, the recognised text is garbage. Fix orientation here first, then send the corrected file to OCR so the text layer is right.
Step 1 Rotate tool: 180° on the flipped pages Step 2 Download corrected PDF Step 3 Open /pdf-tools/pdf-ocr and run OCR Result: searchable text layer aligned to upright pages
Edge cases and what actually happens
You ran the 180° pass twice on the same pages
Flipped backRotation is additive: each run adds your angle to the page's current rotation. Two 180° passes on the same page total 360°, which is the same as 0° — so the page returns to upside down. Apply the correction once; if you double-applied, run one more 180° pass on those pages to fix it.
You typed a hyphen range like 3-7 in the Pages field
By designThe Rotate tool splits the Pages field on commas and reads each token as a whole number, so 3-7 is read as 3 and only page 3 rotates. This is intentional in the rotate path (other tools that need ranges, like delete-pages and extract-pages, parse them — rotate does not). List each page: 3, 4, 5, 6, 7.
A page number is larger than the document length
errorIf you ask for page 99 in a 10-page scan, the engine tries to fetch a page index that doesn't exist and the job fails. The page count shown next to the filename tells you the valid range (1 to N). Re-enter numbers within that range.
You entered page 0 or a negative number
errorPage numbers are 1-based. Page 0 maps to index -1, which is out of range and errors. Use 1 for the first page.
The scan is larger than the free 2 MB limit
Free limitScans of many pages or high-DPI captures easily exceed 2 MB. On the free tier the job is blocked with a notice that free handles files up to 2 MB; Pro raises the limit to 50 MB. To stay on free, shrink the scan first with Compress (lossy).
The scan has more than 50 pages on the free tier
Free limitFree tier caps PDFs at 50 pages; a longer scan is blocked with a page-count notice. Pro raises the cap to 500 pages. Alternatively, split the scan into ≤50-page chunks with the Split tool, rotate each, and re-merge.
OCR text was embedded at the old (upside-down) angle
Re-run OCRRotating the page does not rewrite an existing text layer. If the scan was already OCR'd while upside down, the visible page is now correct but the underlying text is still mis-angled. Re-run OCR after rotating to regenerate an aligned text layer.
The output looks identical to the input
Check angle/pagesIf nothing visibly changed, the most likely causes are: you chose 90°/270° instead of 180°, or your Pages list didn't match the flipped pages (e.g. off-by-one, or a hyphen range collapsed to one page). Confirm the angle is 180° and the page numbers are 1-based and comma-separated.
Page is upside down AND mirrored
Out of scopeRotation only turns the page; it cannot flip (mirror) a page. A genuinely mirrored scan (text reads as a reflection) cannot be fixed by any rotation angle — that requires re-scanning the original the right way round.
Frequently asked questions
What angle fixes an upside-down page?
180°. Pick 180° in the Rotation angle dropdown. A 180° turn takes a page that reads bottom-to-top (header at the foot of the sheet) back to upright. The 90° clockwise and 270° options are for pages that are sideways, not flipped.
Can I fix only some pages and leave the rest alone?
Yes — that is the normal use. Type the upside-down page numbers comma-separated in the Pages field, e.g. 3, 4, 7. Only those pages are rotated; every other page is left exactly as it was. Use all only when the whole scan is flipped.
Does rotating change image quality or make the file bigger?
No. Rotation only updates each page's orientation flag — the underlying scanned image is reused untouched. The output file size and visual quality are essentially the same as the input.
Is my scanned document uploaded anywhere?
No. Rotation runs in your browser via pdf-lib. The scan — passport, statement, medical record, whatever it is — never leaves your device. Only an anonymous usage counter is recorded when you're signed in, with no file content.
Why does running it twice put the pages back upside down?
The rotation is additive: each run adds your chosen angle to the page's current rotation. Two 180° passes total 360°, which is the same as no rotation. Apply the fix once. If you accidentally doubled it, do one more 180° pass on those pages.
Can I type a range like 3-7 to rotate several pages?
No — the Rotate tool does not parse hyphen ranges. 3-7 is read as just page 3. List each page comma-separated: 3, 4, 5, 6, 7. (Range syntax works in the Delete-pages and Extract-pages tools, but not in Rotate.)
What happens if I enter a page number that doesn't exist?
The job errors. The page count is shown next to the filename after the PDF loads — keep your numbers between 1 and that count. There is no silent skip for out-of-range pages in the rotate path.
Will rotating break the OCR text layer?
Rotating doesn't delete the text layer, but if the scan was OCR'd while upside down, the text is still aligned to the old angle. Fix the orientation here first, then re-run OCR so the searchable text matches the upright page.
My scan is over 2 MB — can I still rotate it?
On the free tier, no — files are capped at 2 MB. Either upgrade to Pro (50 MB cap) or shrink the scan first with Compress (lossy), which re-encodes the page images at a lower quality to bring the size down.
Can I rotate a password-protected scan?
If the PDF is readable (e.g. it has owner-level restrictions but opens without a password), yes — the engine loads it ignoring encryption flags. If it requires a password to open, remove that first with Unlock or Remove password, then rotate.
Does this work on scans from any scanner or app?
Yes. The orientation is stored in the standard PDF /Rotate page entry, so the result is honoured by Acrobat, Preview, Chrome, and every conformant viewer — regardless of whether the scan came from a flatbed, an ADF, or a phone app.
What if a page is upside down and also mirror-flipped?
Rotation can't fix mirroring — it only turns the page. A truly mirrored scan (text reads as a reflection) needs to be re-scanned the right way round; no rotation angle will correct it.
Privacy first
All PDF processing runs locally in your browser using PDF-lib and pdf.js. No file is ever uploaded — only metadata counters are saved for signed-in dashboard stats.