This free online document scanner turns photos of paper into a clean PDF: scan to PDF with your phone camera or webcam, no app to install. Take a picture of each page, and the page edges are detected automatically; drag the corners if they are off, and the page is straightened with a perspective correction, so a sheet photographed at an angle comes out as a flat rectangle. Choose the black & white document filter to remove shadows and make text crisp, or keep colour, then reorder, rotate or delete pages and save them all as one PDF on A4 or Letter paper.
Everything happens in your browser. The photos are not uploaded, which matters for the things people usually scan: IDs, contracts, medical forms, receipts and tax papers. The edge detection is deliberately simple, and works best with a page on a darker plain surface; white paper on a white desk, or a page in heavy shadow, may need the corners placed by hand, and the tool says when it could not find them. It makes image PDFs: to get selectable text, run the pages through an OCR tool.
How to use
- Press Take a photo (on a phone this opens the camera) or Choose images. On a computer, Use live camera keeps a webcam view open for page after page.
- Check the four corners on the photo. Drag any that are off, or select one and use the arrow keys; Detect edges again or Use the whole photo reset them.
- Pick a filter (black & white document is best for text) and rotate the page if needed. The preview shows the flattened result.
- Repeat for every page. Use Move up, Move down and Delete in the page list to put them in order.
- Choose the page size, resolution and JPEG quality, name the file, and press Create PDF. It downloads straight away.
Frequently asked questions
How do I scan a document to PDF with my phone?
Open this page in your phone's browser and press Take a photo, which opens the camera. Lay the page on a darker surface, fill the frame, and shoot from straight above if you can. The page edges are found automatically; adjust the corners if needed, choose the black & white filter for a text page, add the next page, and press Create PDF when you are done.
Are my scans uploaded anywhere?
No. The photos are decoded, cropped, straightened and filtered in your browser, and pdf-lib builds the PDF there too. Nothing is sent to a server and there is no account. Closing the tab discards everything.
Why were the page edges not detected?
Detection looks for a bright sheet against a darker background. It fails when the page and the surface are both light, when a shadow cuts across the page, or when the page is not fully in the photo. You can always drag the four corners onto the page corners yourself; a magnifier appears while you drag so the corner is not hidden under your finger.
How big will the PDF be?
At the default 150 dpi and 75% JPEG quality, a page is typically 150 to 500 KB, so a 10-page document is a few megabytes. Black & white pages are the smallest. Lower the quality or resolution for email attachment limits, or raise the resolution to 300 dpi for fine print or archiving. The PDF compressor can shrink it further.
Is the text in the PDF searchable?
No. The scanner produces an image of each page, like a flatbed scanner without OCR. To get text you can search or copy, run the scanned images through the image to text (OCR) tool, which also works in your browser.
Getting a good scan from a phone photo
• Put the page on a dark, plain surface: a desk, a book cover, a jacket. Contrast is what the edge detection looks for.
• Use even light. Daylight from a window beats an overhead lamp, which puts the shadow of your phone on the page. The black & white filter removes gentle shadows, not hard ones.
• Hold the phone above the page rather than at a steep angle. Perspective correction straightens a tilted shot, but the far edge of the page has fewer pixels and comes out softer.
• Let the camera focus before you shoot, and keep the whole page in the frame with a little margin around it.
How the scanner straightens a page
A photo of a flat page taken at an angle is a perspective projection: the edges that are parallel on paper converge in the picture. Four corner points are enough to undo that exactly. The tool solves for the 3 × 3 homography that maps the page rectangle onto the four corners, then builds every pixel of the flat page by looking up where it falls in the photo and blending the four nearest photo pixels.
The corners come from a lightweight detector written for this page: the photo is shrunk to 320 pixels, blurred so the text melts into the paper, split into light and dark with Otsu's threshold, and the largest light region's outline is reduced to its biggest four-sided shape. It takes a few milliseconds and needs no download, which is why it was chosen over bundling the 8 MB OpenCV library; the trade-off is that it needs contrast between paper and background, so the corners are always yours to adjust.
Which filter to use
• Black & white document: each pixel is compared with the average brightness around it, so text stays black and paper turns white even where the lighting fades across the page. Best for letters, forms and printed text, and gives the smallest files.
• Enhanced colour: stretches the contrast so the paper is white and ink is dark while keeping colours. Good for documents with stamps, highlighter or colour logos.
• Greyscale: removes colour casts from warm indoor light without the hard edges of black & white; good for pencil and photographs.
• Original colour: the straightened photo, unchanged.
