Redact a Document

Black out personal data in a screenshot, photo or scanned PDF. The covered pixels are replaced, not hidden — and the file never leaves your device.

Black out personal data

Reads the text in a screenshot, photo or scan, finds things like email addresses and card numbers, and covers them. Check what it found before you apply it — the covered pixels are replaced, so the result cannot be un-redacted.

Covering is not the same as removing

The most common way to redact a document is to draw a black rectangle over the sensitive part and save it. In a PDF that does almost nothing: the rectangle is one object and the text is another, still there in the file, still selectable, still searchable. Courts, government departments and newspapers have all published documents redacted this way and had the text pulled back out within minutes of release.

This tool takes the slower route on purpose. Each page is rendered to an image, the boxes are painted into that image, and the file you download is built from the painted pixels. There is no layer to peel back, because the original was never copied into the output.

The trade-off is that a PDF comes back flattened — page images rather than selectable text. For a scan or a screenshot that costs nothing, since there was no text layer to start with. For a born-digital document it is a real loss, and worth knowing before you start.

Detection is a starting point, not the finish. It reads what it can and matches known shapes; names, addresses, handwriting and anything the recognition misread will not be found. That is why nothing is applied until you have looked at it, and why you can draw over anything yourself.

Redaction FAQ

Can the black boxes be removed from the file afterwards?

No, and this is the part most redaction tools get wrong. A rectangle drawn over a PDF's text leaves the words sitting underneath in the file — the document looks censored and anyone can copy the text straight out of it, which is how redaction failures end up in the news. This tool renders the page to pixels, paints the boxes into those pixels, and builds the output from the result. The original characters are not carried forward, so there is nothing underneath to recover.

Is my document uploaded to be scanned for personal data?

No. The text recognition runs on your own device as WebAssembly your browser downloads once, and the file never leaves the tab. This matters more here than almost anywhere else on the site: the documents worth redacting are exactly the ones you should not be handing to a stranger's server in order to have them censored.

Will it find everything?

No, and you should not use it as though it will. It finds text that matches known shapes — email addresses, phone numbers, card numbers that pass a checksum, IBANs — in text it managed to read. Faint scans, unusual formats, handwriting and names of people all fall outside that. Everything found is shown to you before anything is applied, and you can draw over anything it missed. Treat it as a fast first pass that you check, not an automatic guarantee.

Why does my PDF come back without selectable text?

Because that is what makes the redaction real. Keeping the text layer would mean keeping the words you asked to remove. The output is a PDF of page images, so the redacted content is gone from the file entirely. For the scans and screenshots this tool is aimed at there was no selectable text to begin with; if your PDF is a text document and you need it to stay searchable afterwards, this is the wrong tool for it.

Why are long numbers not detected by default?

Because that rule catches invoice totals, dates, page counts and order references as readily as account numbers, and a preview full of boxes over ordinary figures is one people stop reading. It is there as a toggle for when you want it. Card numbers are checked against the Luhn checksum, which is why they can be on by default without the same noise — roughly nine in ten random sixteen-digit numbers fail it.

Does it remove metadata too?

The output is built from scratch out of the redacted images, so the original file's metadata does not travel with it. If you want to be thorough about a photo specifically, the EXIF stripper is the tool aimed squarely at that — location and camera data are a separate leak from anything visible in the picture.