Document Scanner
Finds the four corners of a photographed document and straightens it into a clean scan.
0 B uploaded — this page has not sent your files to any server?
Turn a crooked phone photo of a document into a clean, flat scan
Without a scanner, a document photographed on a desk tends to come out as a slanted trapezoid with shadows across it. This tool finds the four corners of the paper, flattens it into a head-on rectangle, and can wipe out the shadows with B&W document mode. It handles several photos at once, and the files are processed only on your computer.
A closer look
The scanner runs on OpenCV, an open-source image processing library, and downloads about 10 MB once, on the first scan. The outline is found on a copy of the photo scaled to 1000px on its long side: edges are extracted, the largest convex shape with four corners is treated as the document, and quadrilaterals covering less than 12% of the photo are ignored. The perspective is then corrected using those four points, with processing capped at 3000px on the long side. There is no way to drag the corners by hand, so if no outline is found the photo is saved at its original proportions, without perspective correction, with only your chosen mode applied.
There are three output modes. Color leaves the colors alone and Grayscale simply removes them. B&W document compares each point with the brightness around it to separate ink from paper, so shadows falling across the page and stains disappear, the result looks like a photocopier scan, and the file gets much smaller. The trade-off is that photos or colored stamps on the page break up into dots, so Color or Grayscale suits those documents better. Every result is a JPG at 92% quality, and there is no option to combine pages into a PDF.
How to use it
- Drop photos of documents or receipts on the page or use Choose files. You can add several at once; JPG, PNG, WebP and BMP are accepted.
- Under Output mode, choose Color, B&W document or Grayscale.
- Press Scan. The first time, the scanner downloads, then each photo is processed in turn. Every row gets a thumbnail and either “Corners detected · perspective corrected” or a note that correction wasn’t possible.
- Press Save on a row, or Download all when there are two or more. File names end in “-scan”, and a single scan downloads automatically.
When it helps
Common snags
Tip Even when the corners are found, the steeper the angle, the more the far side of the page is stretched and softened, so shooting from directly above gives the cleanest result.
Document Scanner FAQ
How should I take the photo?
Strong contrast between the document and the background works best. Place a white document on a dark desk, keep all four corners in frame, and let the document fill most of the shot.
What if the corners aren't detected?
If no document outline is found, the image is processed without perspective correction, at its original proportions, with only your chosen mode (color, B&W, grayscale) applied. Try again with a higher-contrast background.
What does B&W document mode do?
It removes paper shadows and stains and turns the page into crisp black-and-white text, like a photocopier scan — great for submitting paperwork, and the file gets much smaller. For documents with photos, use color or grayscale.
Can I scan iPhone HEIC photos?
Convert them with HEIC to JPG first, then scan the JPG.
Can I extract the text afterwards?
Save the scan, then run it through Image to Text (OCR). B&W document mode improves recognition accuracy.
Is it free? Are my files safe?
Completely free with no limits. Everything runs in your browser and your photos are never uploaded to any server.