Redact a PDF Properly

A black rectangle is not redaction. Drawing a box over a name leaves the name in the file, where anyone can select, search or extract it. This page deletes the characters themselves, then reopens the copy it made and looks for them again.

The tool runs in your browser (JavaScript required). Nothing is uploaded.

Why a black box is not redaction

Most ways of hiding text in a PDF only add something on top of it. The drawing sits in one layer, the text stays in another. Copy the page into a text editor, or open the file with a different reader, and the words are still there — this is how redaction failures end up in the news.

Here the glyphs are removed from the content of the page itself and the space they took is closed up, so the rest of the line stays where it was. A box and a label are drawn afterwards only so a reader can see that something was taken out.

Selecting, searching or copying the page no longer finds the value
The same value is also removed where it is stored away from the page: comments, form values, bookmarks, document properties and XMP
Earlier saves kept inside the file are dropped, so the old text cannot be recovered from them
You choose per item: delete it, replace it with a label like [PERSON_1], mask it, or keep it

How the result is checked

The copy is written first and judged afterwards. It is reopened with a second, independent reader, every string and every readable stream inside it is searched, and the detectors run again from scratch. Only if none of them finds the value does the result say it was removed.

When part of the file cannot be read — a scanned page with no text layer, an unusual font encoding — the result says "Could not fully verify" instead of claiming success. Nothing is uploaded: the file is read, changed and checked inside this browser tab.

Questions

Can the redacted text be recovered from the new copy?

Not from the page content: the characters are deleted, not covered, and the file is written again from scratch so earlier saves are not carried over. We show you the check: the new copy is reopened and searched for every value you chose to remove.

Does the layout shift after redaction?

No. The deleted glyphs are replaced by the exact space they used, so the rest of the line and the following lines stay where they were.

What about text inside a scan or a picture?

It is not read. This version has no OCR, so a page that is a picture is reported as not fully checked — we never treat "no text found" as "nothing to find".

Can I redact something the detectors don't know about?

Yes. Type it into "Add something you don't want to share" — a name, a case number, a price. It is matched in the page text and everywhere text is stored in the file.

Is the file sent to a server?

No. Everything runs in your browser tab. You can disconnect from the network after the page has loaded and it still works.