A PDF page straightener corrects tilted or rotated pages in scanned PDF documents. Use manual rotation for precise fine-tuning (±15° in 0.5° steps) or auto-detect for batches of standard-rotation issues. All processing happens locally — no file uploads.
Upload PDF
Drop a scanned PDF here or
Best for scanned PDFs with tilted or rotated pages
document.pdf
Auto-detect uses Tesseract.js OCR
Downloads ~6MB on first use (cached afterward). Works best on text-heavy pages. For pages with primarily images or blank content, use Manual mode.
How to Straighten Tilted PDF Pages
Scanned documents frequently end up slightly tilted — the paper wasn't perfectly aligned in the scanner, or the original document had physical creases or bends. This PDF deskew tool corrects this in two ways: fine manual rotation using a slider, or automatic detection using Tesseract.js orientation analysis.
Manual Rotation Mode
Best for precise fine-tuning. Use the slider to apply rotation between -15° and +15° in 0.5° increments — this handles typical scan tilt (most scans are off by ±1° to ±3°). For major orientation errors (pages scanned sideways or upside down), use the preset buttons: +90°, 180°, -90°. You can apply to all pages or a specific page range.
Auto-Detect Mode
Best for batches of pages with standard orientation issues. Click "Auto-Detect & Straighten" to load Tesseract.js (~6MB, downloaded once and cached). The tool renders each page with PDF.js, runs orientation detection, and applies the detected rotation using pdf-lib. Auto-detect identifies rotations at 90° increments (0°, 90°, 180°, 270°).
Tips for Best Results
Auto-detect works best on text-heavy pages with clear, printed characters. For pages with diagrams, photos, or blank pages, auto-detect may not give useful results — use manual mode instead. For sub-degree tilt correction (the most common scan issue), the manual slider gives the most control.
FAQ
How does automatic deskew work?
The auto-detect mode renders each page using PDF.js, then runs Tesseract.js orientation detection on the rendered image. Tesseract analyzes the text angles and returns a detected rotation (0°, 90°, 180°, or 270°). The tool then applies the inverse rotation using pdf-lib to straighten the page.
How accurate is automatic detection?
Auto-detect works well for text-heavy pages with clear, printed text. Accuracy decreases for blank pages, pages with primarily images, heavily rotated text, or handwriting. For best results, use manual adjustment on pages where auto-detect doesn't work well.
What's the difference between manual and auto deskew?
Manual deskew lets you apply a specific rotation angle (-15° to +15° in 0.5° steps, plus 90°/180°/270° for large corrections) to selected pages. Auto-detect uses Tesseract.js OCR to guess the rotation and applies it automatically — useful for batches of scanned pages.
Why does auto-detect only detect 0, 90, 180, or 270 degrees?
Tesseract.js orientation detection works at 90° increments (the four standard page orientations). For fine-grained deskewing of slightly tilted scans (±0.5° to ±15°), use the manual slider which gives precise control in 0.5° steps.
Is my PDF file uploaded to any server?
No. All processing runs in your browser using PDF.js for rendering, Tesseract.js for orientation detection, and pdf-lib for applying rotations. Nothing is uploaded anywhere.