Document Scanner

A photo of a document taken at an angle, on a patterned desk, under uneven lighting, never quite looks like a real scan — the page is skewed into a trapezoid, the background is visible around the edges, and the lighting makes parts of the page look gray or yellow instead of white. This tool fixes all of that directly in your browser: upload a photo, rotate it if it came out sideways, drag the four corner handles onto the actual edges of your document, and it straightens the perspective, cleans up the background lighting, sharpens the contrast, and lets you download the result as a PDF — the same kind of output you'd get from a dedicated scanning app, without installing one or uploading anything.

Perspective correction is the core of the tool. Because a photographed page is almost never perfectly parallel to the camera, its four corners rarely form a true rectangle in the image — they form some skewed quadrilateral. After you mark those four corners, the tool computes a projective transform that maps that quadrilateral back onto a straight rectangle, resampling every pixel of the output from the original photo. This is the same category of math used by professional scanning apps to "de-skew" a photographed page, run entirely as plain JavaScript against the pixels already sitting in your browser's memory.

Once the page is straightened, two more corrections happen automatically. Background cleanup estimates the photo's uneven lighting (the shadows and hotspots a phone camera almost always introduces) and divides it out, so a page that looked patchy or gray becomes an even, bright background. Contrast optimization then stretches the resulting brightness range so the page's whites and blacks use the full range instead of a washed-out middle band — the same idea as an "auto levels" adjustment. You can turn this enhancement off if you'd rather keep the corrected-but-unprocessed colors of the original photo.

Finally, choose Color to keep the page's natural colors (good for photos, colored diagrams, or documents where color carries meaning) or Black & White to convert it to a crisp, high-contrast grayscale scan (good for plain text documents, where it also tends to produce a smaller, cleaner-looking result). The finished page downloads as a single-page PDF sized to match the corrected document's proportions — ready to attach, print, or file away. Nothing about your photo — the image data, the corners you drag, or the finished scan — is ever uploaded, logged, or sent anywhere; every step runs locally using the Canvas API already built into your browser.