Document Scanner — Photo to Clean PDF

A photograph of a document is not a scan. Held in one hand, the page comes out tilted, the corners are trapezoids rather than right angles, one side is brighter than the other, and the shadow of your own phone falls across the middle. Emailed to an office, that photo often gets rejected. This tool turns it into something that looks like it came off a flatbed.

Point the camera at the page, or drop in photos you have already taken. The tool looks for the four straight lines that bound the paper, works out where the corners are, and shows them to you as draggable handles so you can correct anything it got wrong. It then applies a perspective transform — mathematically the same correction a copy-stand would give you — so the trapezoid becomes a true rectangle with the text running straight across.

The clean-up stage is what makes it readable. Rather than one brightness threshold for the whole image, which blows out the bright side of the page and loses the dark side, each pixel is compared against the average of its own neighbourhood. Uneven lighting stops mattering, and you get solid black text on genuinely white paper. There are gentler options too: greyscale keeps shading for photographs and signatures, boost keeps colour while removing the yellow cast of indoor lighting, and original just straightens without touching the tones.

Add as many pages as you need, reorder them, rotate any that came out sideways, and export the lot as a single multi-page PDF at A4, US Letter or the image's own proportions. All of it — the edge detection, the perspective maths, the thresholding and the PDF itself — runs inside your browser on your own device. Your documents are never uploaded, which matters rather a lot when the thing you are scanning is an ID card, a bank statement or a signed contract.