Offline Kit
한국어
PDF redaction

A black box doesn't redact anything

Cover text with a filled rectangle in Preview, Acrobat, or any free online editor, and the words stay in the file underneath. Anyone can select them, copy them, or pull them out with a script. This has leaked court filings, medical records, and salary figures — repeatedly, from people who believed the document was safe.

See it happen

Both documents below were built in your browser a moment ago. The left one has black rectangles drawn over the numbers. The right one was rebuilt from re-rendered pixels. Underneath each, the text extracted straight from the resulting file.

Rectangle drawn on top

checking
NameHong Gil-dong
ID number880101-1234567
Phone010-2847-9931
Text extracted from the file
running…

Page rebuilt from pixels

checking
NameHong Gil-dong
ID number880101-1234567
Phone010-2847-9931
Text extracted from the file
running…
01

Check a document you already sent

This takes about ten seconds and needs nothing but the PDF reader you already have.

  1. Open the PDF and select the whole page — ⌘A on macOS, Ctrl+A on Windows.
  2. Copy, then paste into any plain text editor.
  3. Read what came out. If the text you meant to hide appears, the redaction never happened.

A second check, for the careful: some readers refuse to select text under a shape. Run the file through any "PDF to text" converter instead. If the words are in the file, a converter will find them.

02

Three ways to remove it for real

All three work. They differ in what you give up.

Print the page to an image, then back to PDF

Export each page as PNG or JPEG, cover the regions in any image editor, and reassemble. The characters are gone because the page is no longer made of characters.

Free, works anywhere, and tedious past a couple of pages. Easy to miss a page.

Acrobat Pro's redaction tool

Not the rectangle drawing tool — the Redact tool, which deletes the underlying content and then flattens the page. Mark the regions, then Apply.

Reliable when you use the right tool and actually click Apply. Subscription, and the file goes through Adobe's software on your machine.

Re-render the page, then paint over it

Rasterise each page at print resolution, fill the regions, and rebuild the PDF from those pixels. This is what the tool below does, in the browser, with no upload.

The output is an image, so the remaining text is no longer selectable or searchable. That is the price of removing it properly.

03

The text isn't the only thing that leaks

Even a correctly redacted PDF can carry things you did not mean to send:

Rasterising the page removes all of these at once, because the rebuilt file contains only images. That is a side benefit of the third method, and a reason to prefer it for anything going to a court, a regulator, or a stranger.

Does auto-detection catch everything?

No, and no tool's does. Pattern matching finds numbers in the shape it was taught — an ID number written with spaces instead of a hyphen, a phone number split across a line break, or a name that only a human recognises as sensitive will slip past. Treat it as a first pass that saves you time, then read the page yourself before you export. The one thing you should never do is export without looking.

A tool that does the third method

Redact runs as a single HTML file in your browser. It finds ID, card, phone and email patterns across the document, you check them, and it exports a rebuilt PDF. Nothing is uploaded — turn off your network and it still works.

See Offline Kit