Data Pantry logo 데이터팬트리

Document Scanner

Scanner-quality results without a scanner. Automatically corrects perspective distortion and shadows in your smartphone photos to produce clean, flat document images.

Click or drop an image

JPG · PNG · WEBP · Up to 50MB · Multiple files supported · Paste with Ctrl+V

Upload an image

Upload a photo of your document and we'll automatically detect its edges and flatten it like a scan.

User guide

What it does

Automatically corrects the perspective distortion in a photo of a document taken with your smartphone, turning it into a neatly aligned image as if it had been scanned. Supports shadow removal, color mode selection, and multi-page ZIP export.

How to use
  1. Upload an image: Drag and drop a photo of your document onto the drop zone.
  2. Adjust corners: The document's four corners are detected automatically (even for tilted documents). If the paper is curled or wrinkled so its edges aren't straight, that curvature is detected too and flattened automatically. If detection isn't accurate, drag the handles to adjust it yourself.
  3. Rotate: If the photo was taken sideways, use the left/right 90° buttons to fix the orientation.
  4. Choose a color mode: Select B&W, grayscale, or color mode.
  5. Brightness, contrast and shadow removal: Adjust brightness and contrast with the sliders, and shadows from lighting are removed automatically.
  6. Click Scan to download a perspective-corrected image.
When photographing your document, make sure the whole document is within the frame — a solid-colored background around it improves auto-detection accuracy.
Color modes
  • B&W: Best for text documents. Makes the background white and the text black for clarity.
  • Grayscale: Suited to pencil sketches and handwritten notes.
  • Color: Use for documents containing photos or when color information matters.
Multi-page handling

When you scan multiple pages, you can save each image as an individual PNG, or bundle them into a ZIP file to download all at once.

Limitations and notes
  • If the document's background is similar in color to the paper, or if lighting is very bright on one side and very dark on the other, automatic corner/curve detection may be inaccurate. In that case, drag the corner points manually or click along an edge to add bend points yourself.
  • OCR text recognition works by downloading a language pack from an external CDN (Tesseract.js). If a corporate or institutional network blocks external CDNs, only OCR will fail — scanning and downloading will continue to work normally.
  • OCR is optimized for printed text, so handwriting or low-resolution photos may have lower recognition accuracy.
  • Each file can be up to 50MB, and up to 30 images can be queued for batch processing.
  • All processing happens in your browser, so very high original resolutions or a large number of pages may process more slowly depending on your device's performance.
FAQ
The corners aren't detected automatically.

Just drag the blue corner points to the document's four corners yourself.

I converted to B&W but the text isn't very legible.

Adjusting the threshold slider can sharpen the line between text and background.