Is the text really gone?

Yes. Matching text is removed from the document's content rather than covered over, so there is nothing beneath the marking to select, copy or extract. Verify by searching the result for the terms.

Why is drawing a black box not enough?

Because the box is a graphic on top and the text is still in the file. Selecting it, copying it or extracting the content recovers it in seconds. This is the mistake behind a long series of public disclosures.

Will it work on a scanned document?

No. A scan contains images of text, so there is nothing for a search to match. Use the image blur tool to destroy those regions in the pixels, which is the equivalent operation for images.

Redacting the body does not touch the author, title, creation software or revision history, and those are a well-known route to what a document previously said. Clear them separately before releasing anything sensitive.

Redact PDF

Permanently remove words from a document, not just cover them.

Everything runs inside your browser. Your files never leave your device.

Drop your file here

.pdf · they stay on this device

Ask about Redact PDF

Questions about what this tool does, which option to pick, or what it can and cannot handle.

The question you type here is sent to an AI provider to be answered — your files and whatever you put in the tool above are not, and the assistant cannot see them. Answers are generated and can be wrong. The tool itself is not guessing: it runs deterministic code on your device.

About the Redact PDF

Redaction means removing information from a document, and the distinction between removing and concealing is the entire subject. Drawing a black rectangle over a name in a PDF hides it from view and leaves the text underneath, where anyone can select it, copy it, or extract it with a script. Public bodies, law firms, courts and newspapers have all released documents that way and all had the covered text read back to them.

This does the other thing. Name the words or phrases to remove, and every occurrence is taken out of the document's text and covered. The text is not there to recover, because it has been removed rather than obscured.

Search is across the whole document rather than page by page, which matters because the thing you are redacting is rarely in one place. A name appears in a header, a signature block, a footnote and the body; missing one occurrence undoes the entire exercise, and that is the ordinary way redaction fails.

Two limits stated plainly. Scanned documents contain images of text rather than text, so there is nothing for a search to find — those need the region blurred as an image, which the image blur tool on this site does destructively. And metadata is separate from page content: the author field, the title and the revision history are not touched by redacting the body, and they are a well-known route to what a document used to say.

Check the result by searching it for the terms you removed. If the search finds nothing, the redaction worked. That check takes ten seconds and is the difference between a redacted document and an incident.

How to use it

1Drop in the PDF.
2List every term to remove, one per line — including the variants: full name, surname alone, initials, email address.
3Produce the redacted document.
4Open it and search for each term. Finding nothing is the confirmation; skipping this step is how redaction fails.

Questions

Is the text really gone?: Yes. Matching text is removed from the document's content rather than covered over, so there is nothing beneath the marking to select, copy or extract. Verify by searching the result for the terms.
Why is drawing a black box not enough?: Because the box is a graphic on top and the text is still in the file. Selecting it, copying it or extracting the content recovers it in seconds. This is the mistake behind a long series of public disclosures.
Will it work on a scanned document?: No. A scan contains images of text, so there is nothing for a search to match. Use the image blur tool to destroy those regions in the pixels, which is the equivalent operation for images.
What about metadata?: Redacting the body does not touch the author, title, creation software or revision history, and those are a well-known route to what a document previously said. Clear them separately before releasing anything sensitive.