ScanHome › Clean a scanned document
Clean a scanned document online
Your file is deleted the moment it's done.The big PDF sites keep your bank statements and IDs on their servers. Here the original goes right after processing, the result after 45 minutes — or sooner, on your click.
A phone photo of a document is rarely tidy: it's tilted, the edges of the desk show, and one side is darker than the other. ScanReviver cleans that up in a few seconds: crop to the page, straighten it, even out the lighting, and save a neat PDF or image. It's free, needs no account, and your upload is deleted right after processing.
The edges are found for you here. Want to drag the corners yourself, reorder pages or switch between colour and black & white? Open the full scanner.
Clean your document now
Upload a JPG, PNG, WebP, HEIC, TIFF, BMP, GIF or AVIF photo: no sign-up, nothing to install.
Open the tool →What we see before a document even gets cleaned
Not every upload makes it to the cleaning step. Over the 38 days we've been tracking (15 July – 21 August 2026), 23 uploads were turned away before ScanReviver touched them at all — compared with zero jobs that started cleaning and then failed partway through. That gap matters: a document that gets rejected never gets processed, so there's nothing left over to clean up or store; a document that's accepted reliably finishes. The traffic behind those numbers is smaller than a big PDF site's but genuinely spread out: 3,950 page views from 257 real visitors (bots excluded) across 27 countries in that window. The US (30) and China (20) are the largest single groups, but the same page has also been opened from Germany, France, India, Egypt, Brazil and Vietnam — a reminder that a crooked phone photo of a document is a universal problem, not a regional one, which is part of why this tool doesn't ask for a country, a language or an account before it works.
What "cleaning" a document actually does
ScanReviver runs a small, predictable set of steps on your image:
- Find the page: it detects the document's four corners, and you can drag them to fine-tune the crop so the desk, hands and background are cut away.
- Straighten it: perspective correction squares up a page shot at an angle, and small tilts are levelled so text runs horizontally.
- Even out the lighting: it lifts the background toward white and reduces uneven shadows, so a page that was grey and blotchy reads clearly.
- Choose a look: keep it in colour, produce a clean near-white document, or go crisp black & white for text-only pages.
- Export: download a JPG, PNG, or a PDF (single- or multi-page), optionally with a searchable text layer.
Colour, clean or black & white?
Pick the mode that matches the page:
- Colour Document: keeps colours (letterheads, forms with coloured fields, highlighted notes) while still flattening the lighting.
- Clean Document: pushes the background to near-white for a scanned-in look; the best all-round choice for printed text.
- Black & White: high-contrast, smallest files, ideal for plain text you plan to print or archive.
What the cleanup assumes about your page
The Clean and Colour modes work by estimating the paper background across the whole image and dividing the photo by that estimate. That is what flattens uneven lighting while keeping the ink dark. It rests on one assumption: the paper is the lightest thing in the frame.
Most of the time that holds, and the result looks like a real scan. Two situations break it, and they're worth recognising:
- A mostly dark image — a photo of a screen, a slide with a black background, a page shot in near-darkness. There the estimate would latch onto a few bright pixels and push everything else towards white unevenly, channel by channel, which shows up as a colour cast. The tool checks for this and skips the correction on a dark image rather than applying an assumption that doesn't hold. If your result looks untouched, that's usually why: re-shoot with more light.
- A big dark photograph on the page. A full-width picture in a magazine article counts as "not paper" over a large area, and the correction can lighten locally around it. Cropping tightly to the text you actually need avoids it.
Cleaning is not compressing
Cleanup changes how a page looks; the mode you pick also decides how big the file is, which matters when a portal caps uploads at a few megabytes.
- Colour keeps all three channels and the most detail: the largest file.
- Clean pushes the paper to near-white. Large flat white areas compress well, so the file is typically noticeably smaller than Colour.
- Black & White is a hard threshold, so the page becomes two tones. Smallest by a wide margin, and the right choice for a text-only archive — as long as nothing on the page is faint, because a threshold either keeps a pixel or throws it away.
If the goal is "as small as possible while still legible", try Black & White first and check the smallest text on the page at full zoom before you commit. If a faint line has vanished, step back to Clean.
Frequently asked questions
Is it really free?
Yes. Cleaning documents is free and needs no account. Ads on these pages help cover the server costs.
What happens to my files?
Your original upload is deleted right after processing. Your cleaned result stays available for a short time so you can download it, then it's deleted automatically, and you can delete it immediately with the "Delete now" button. Files are never used for AI training and are never sold or shared.
Can it clean a multi-page document?
Yes. Add several photos, arrange the order, and download one combined PDF.
Will it change the text on my document?
ScanReviver doesn't use generative AI, so it never rewrites text, changes numbers or recreates signatures or stamps. It does adjust the image (contrast, colour, black & white), so fine details can shift — always check the result before relying on it.
Honest limitations
ScanReviver improves readability; it does not certify or authenticate documents. Very dark shadows, text hidden under objects or fingers, and severe blur may still look imperfect after cleaning. If automatic page detection misses, you place the four corners yourself. It is a cleanup tool, not a way to recover detail that the photo never captured.
Once a scan is clean, enhance scanned document sharpens contrast and text further. If shadows from a phone camera are the real problem, remove shadow from document targets exactly that. Working from an old printed photo rather than a document? The cleanup here happens after the shot is taken — cleaning the physical print first, dust and all, before you scan it gets you a better starting image than fixing it digitally afterwards ever can.