How to Deskew a PDF: 3 Ways to Straighten Crooked Scanned Pages

I usually notice a skewed PDF after OCR goes wrong: the scan looks readable, but the recognized text breaks into odd boxes, missing characters, or slightly shifted lines. The hidden distinction is this: deskewing changes the page image; OCR adds or rebuilds a text layer. If you want a clean editable PDF, straighten the scan first, then run OCR. For the full editing workflow after deskewing, keep this pillar guide open: how to make a PDF editable.

One honest note before the methods: UPDF is not a one-click automatic deskew detector. Its real UPDF workflow is manual: open the scan, switch to Image PDF, and drag the Rotation Angle slider until the text sits level. That makes it strong for sensitive files because the PDF stays on your device. If you need automatic angle detection, use Sejda or Adobe Acrobat instead.

Part 1. Choose the Right Deskew Method

Your situation	Best method	Why
You need to straighten a private scan without uploading it	UPDF manual Rotation Angle slider	Local workflow with exact visual control.
You want automatic angle detection in a browser	Sejda Deskew PDF	Fast for non-sensitive files and mixed batches.
You already pay for Acrobat	Adobe Acrobat Scan & OCR	Built-in Deskew setting inside scanned PDF enhancement.
The page is crooked and you need editable text afterward	Deskew first, then OCR in UPDF	Straight pages produce better OCR and cleaner editing.
You are on iPhone, iPad, or Android	Mobile scan cleanup + OCR	Mobile can rescan, crop, rotate 90 degrees, and OCR, but it does not replace the desktop slider.

Part 2. Deskew vs. Rotate vs. OCR

Action	What it fixes	What it does not fix
Rotate	A page facing sideways or upside down in 90-degree steps	A page leaning by 2 or 3 degrees
Deskew	Small angle tilt across the page image	OCR errors, blur, shadows, or page warping
Crop	Scanner edges, phone background, or black corner triangles after rotation	The text line angle inside the page
OCR	Searchable, selectable, or editable text	The visual angle of the scanned image

Think of rotation as turning a sheet of paper, deskew as nudging a slightly crooked frame until it hangs level, crop as trimming the border, and OCR as placing a transparent text layer over the image. Use them in that order when needed: rotate, deskew, crop, then OCR.

Part 3. How to Deskew a PDF with UPDF on Windows or Mac

Use UPDF when you want full control and do not want to upload the file. UPDF's sibling guide shows a manual deskew route through Image PDF mode and the Rotation Angle slider. Because this is a manual angle adjustment, it is best for files where you can visually judge the page tilt.

Step 1. Open the scanned PDF in UPDF on Windows or Mac. If you need to install it first, download UPDF for free and open a copy of the file.

Step 2. Click the dropdown arrow in the upper-right corner of the UPDF window.

Step 3. Select Image PDF from the dropdown toolbar.

Step 4. Drag the Rotation Angle slider left or right until the text lines look level. Make small adjustments first; most crooked scans are only a few degrees off.

Step 5. Save the corrected PDF. Use Save As Other if you want to preserve the original scan as an untouched backup.

UPDF's advantage here is privacy and control. Automatic deskew tools are convenient, but they can over-correct pages that contain tables, stamps, mixed orientations, or unusual margins. With UPDF, you decide when the page is actually level before moving on to OCR.

Download UPDF for free to try manual PDF deskew on your file. Installation is free, and Pro features are available when you need watermark-free export, batch OCR, or advanced tools.

Part 4. Straightened the Scan? Make It Editable or Searchable with UPDF OCR

Deskewing fixes how the scanned page looks. It does not make the text editable by itself. Once the page is straight, use UPDF OCR to make the file searchable, selectable, or editable. UPDF's OCR supports 38+ languages, and you can select more than one language for a mixed-language scan. This is also the most natural place to continue to the pillar workflow: make the PDF editable.

On Windows or Supported Mac Builds

UPDF's OCR is available on Windows and on Mac with an Apple chip in the official website version. If OCR does not appear on your Mac build, update UPDF from the official site or use the desktop where OCR is available.

Step 1. Open the straightened PDF in UPDF.

Step 2. Go to the left-side Tools panel and choose OCR.

Step 3. Select the document language. If the scan contains more than one supported language, choose each matching language before conversion.

Step 4. Choose the OCR result type:

OCR mode	Use it when
Editable PDF	You need to edit recognized text after OCR.
Searchable PDF Only	You want to preserve the scan's look while adding searchable/selectable text.
Text and Pictures Only	You want a cleaner re-laid-out file and do not need the original scan appearance preserved.

Step 5. Set the page range, click Convert, and open the output file.

Step 6. Test the result by selecting a sentence. If you chose Editable PDF, click Edit in the top toolbar to correct any OCR mistakes.

OCR is recognition, not magic. Deskewing first helps because upright, level text is easier to recognize than a tilted phone photo.

On iPhone, iPad, and Android

UPDF mobile can create, edit, OCR, convert, annotate, and organize PDFs, but do not assume it has the same Rotation Angle slider as desktop. Use mobile for the parts it can do well: rescan, crop, rotate 90-degree page orientation, and OCR.

Step 1. If you still have the paper document, rescan it with the phone parallel to the page. Avoid shooting from one side, because that creates perspective distortion rather than simple skew.

You can also watch the video tutorial below.

Step 2. Before saving the scan, review the capture. Crop away the table or background, rotate the page if it is sideways, and apply a readable filter if needed.

Step 3. If the PDF already exists, import it into UPDF on iPhone, iPad, or Android.

Step 4. Use OCR on mobile when the page is readable enough. For a scan where you care about exact visual fidelity, send the file to desktop and use the UPDF Image PDF slider first.

Step 5. Open the OCR result and proofread names, totals, dates, ID numbers, and contract clauses. OCR quality depends heavily on the input scan.

Mobile matters because many skewed PDFs are created on phones. But the honest distinction stays the same: mobile cleanup can prevent skew, and OCR can rebuild text; the desktop slider is the UPDF method for manually straightening the scanned page image.

Part 5. How to Deskew a PDF Online Automatically with Sejda

Use Sejda when the file is not sensitive and you want the tool to detect the angle automatically. Sejda's Deskew PDF page supports automatic deskew and manual page-by-page adjustment with small angle controls.

Step 1. Open Sejda Deskew PDF.

Step 2. Upload the scanned PDF from your computer, cloud drive, or a web address.

Step 3. Choose Automatic if you want Sejda to detect and straighten skewed pages.

Step 4. Choose Manually page by page if only some pages are skewed or if you want to review each correction. Use the rotation slider or the small angle buttons to fine-tune the result.

Step 5. Click Deskew PDF and download the corrected file.

Step 6. Open the output in UPDF if you still need OCR, editing, annotation, or conversion.

Sejda states that files are automatically deleted after 2 hours, and its free service supports documents up to 200 pages or 100 MB with 3 tasks per hour. That is convenient for ordinary files, but local desktop processing is still the safer default for sensitive documents.

Part 6. How to Deskew a PDF in Adobe Acrobat

Adobe Acrobat can automatically straighten tilted scanned pages through its scanned PDF settings. Use it if you already have Acrobat Pro or your workflow is already built around Adobe.

Step 1. Open the scanned PDF in Adobe Acrobat on desktop.

Step 2. Choose All tools > Scan & OCR.

Step 3. Select Enhance scanned file.

Step 4. Open Enhance scanned file settings. In the Filters section, click Edit and enable Deskew.

Step 5. Adjust related settings if needed, such as background removal, descreen, compression, or text sharpening.

Step 6. Enable Recognize Text if you also want OCR, then choose the language and output type.

Step 7. Click OK, then click Enhance and review the output.

Adobe is useful for automatic deskew, but do not make it the default recommendation for everyone. For private files and occasional manual corrections, UPDF's local slider workflow is lighter; for non-sensitive automatic deskew, Sejda is faster.

Part 7. Method Comparison

Method	Deskew type	Platform	Privacy	Cost	Main limit
UPDF Image PDF + Rotation Angle	Manual micro-angle correction	Windows, Mac desktop	File stays local	Free to try; Pro $49.99/yr or $79.99 one-time	No automatic angle detection; you adjust by eye.
UPDF mobile cleanup + OCR	Rescan/crop/rotate/OCR workflow	iPhone, iPad, Android	File can stay in app workflow	Same UPDF account/license path	Not the same as desktop Image PDF slider.
Sejda Deskew PDF	Automatic or manual page-by-page	Web	Upload required	Free tier	200 pages / 100 MB / 3 tasks per hour; privacy trade-off.
Adobe Acrobat	Automatic Deskew in scan enhancement	Windows, Mac desktop	Local desktop workflow	About US$239.88/year (US$19.99/mo, annual billed monthly)	Requires a paid Acrobat subscription.

If the page image itself must stay faithful to the original, use a deskew method before OCR. If you only need a clean editable result and the original scan appearance does not matter, OCR modes like Text and Pictures Only may be useful, but call that re-layout, not deskew.

UPDF is free to try, but export without a trial watermark and batch OCR require Pro. That is why it is useful to test the manual deskew and OCR flow on a copy first, then upgrade only if you need clean export or heavier OCR work.

Part 8. Edge Cases When Deskew Still Fails

"I used OCR, but the page image is still crooked."

OCR adds or rebuilds text. It does not rotate the original scan image. Use UPDF's Image PDF Rotation Angle slider, Sejda, or Acrobat Deskew first, then run OCR.

"Text and Pictures Only made the page look straight, but the layout changed."

That OCR mode can drop the original scan background and rebuild text/images into a cleaner layout. This can help readability, but it is not layout-preserving deskew. Use Searchable PDF Only after deskewing if the original appearance matters.

"The slider straightened the text but left black triangles in the corners."

Rotating a rectangular page exposes empty background at the corners. Crop the page after deskewing, or keep the corrected copy if the corners do not affect reading or OCR.

"The page looks like a trapezoid, not just tilted."

A trapezoid shape comes from perspective distortion, usually from photographing the page at an angle. A deskew slider can correct line angle, but it cannot fully rebuild the page geometry. Rescan with the phone parallel to the document.

"The file will not save after I straighten it."

The PDF may be password-protected or restricted from editing/organizing pages. Enter the owner password or remove restrictions only if you own the file or are authorized to modify it, then save a corrected copy.

Part 9. FAQ

Conclusion

Deskewing a PDF starts with the right diagnosis. If the page is only sideways, rotate it. If the text leans by a few degrees, deskew it. If you need editable text, run OCR after the page is straight.

For UPDF users, the honest deskew method is the desktop Image PDF workflow with the Rotation Angle slider: manual, local, and precise. After that, use UPDF OCR to make the straightened scan editable or searchable, or continue through the full make a PDF editable workflow. Use Sejda when you want automatic online deskew for non-sensitive files, and use Adobe Acrobat when you already pay for Acrobat's scan enhancement tools.

Download UPDF for free to manually deskew a scanned PDF and then OCR it. Installation is free, and Pro features are available when you need watermark-free export, batch OCR, or advanced tools.