Deskew PDF ：，

Why Scans Come Out Crooked

Sheet-fed scanners grab paper at a slight angle. Phone photos of documents tilt a degree or two no matter how carefully you hold the camera. The result is a PDF where every line of text slopes just enough to look sloppy — and enough to confuse OCR engines that expect horizontal baselines.

Deskewing measures that tilt on each page and rotates the page back. OxygenPDF does it with projection-profile analysis: the tool projects the page's ink onto rotated axes across a ±10° range and finds the angle where text rows line up horizontally. Pages that are already straight are left completely untouched — their vector content and text layer carry over unchanged.

What You Can Control

Auto-detection handles most pages, but the knobs are there for the ones it doesn't.

No Cropping, Ever

Rotating a rectangle inside a same-sized frame clips the corners. This tool pads each rotated page to its new bounding box instead, so the output page grows slightly and every edge of the original survives. Corners fill with white, matching the paper.

The DPI setting (72–300) controls the resolution corrected pages are rendered at. Straight pages are copied through as vectors, so a document where only three pages are skewed keeps its original quality everywhere else.

How It Compares

Deskewing is a checkbox feature in scanner software and a manual chore in editors. Here's where this sits.

Straighten Before You OCR

Skew is the single biggest fixable cause of bad OCR. Text recognition engines segment lines horizontally; a 3° tilt smears characters across line boundaries and accuracy drops fast. Run a crooked scan through Deskew PDF first, then OCR PDF — the searchable output is noticeably cleaner.

Everything runs in your browser. Scanned contracts, medical records, and archive pages never leave your device, which matters for exactly the documents people scan.