How to redact a PDF properly

Redacting a PDF means taking something out of it so that nobody who receives the file can get it back. Most failed redactions are not careless. They are a black box drawn in the right place with a tool that only covered the text, and the text was still there for anybody who selected it.

Open the tool Free below twenty pages, in your browser. Nothing is uploaded.

Before

After

A product line-up before and after: brand names printed on the packaging photographs are covered, as are the names beneath them.

The mistake almost everybody makes once

A black rectangle, a black highlighter or a shape drawn in a PDF viewer is an annotation. It sits on top of the page; the words underneath are still in the file. Select across the bar and paste, and they come out. Search the file for the name, and it is found. Printing to PDF often keeps the text layer too, so it is not a reliable fix.

Proper redaction removes the content itself: the characters, the picture under the bar, and anything else that says what was there. That includes the parts of a PDF you do not see on the page: the document properties, the file name, earlier versions saved inside the file, and the invisible text layer a scanner adds so that a scan can be searched.

The method, whatever tool you use

First, list what has to go: the names, the account numbers, the logos, the signatures. Then find every copy of each, not only the ones you remember. Typed words can be searched for; pictures, logos and words inside scans or screenshots cannot, and that is where most leaks are.

Then apply the redactions so the content is removed, not covered, and save a new copy rather than overwriting the original. Keep the original somewhere safe: once the redaction is applied, it is meant to be impossible to undo.

The fast way: say what, not where

Redaction is slow because it is usually done one mention at a time: find the name, draw a box, scroll, find it again. A forty page document with one client name in it can take an hour, and the mention that gets missed is the one in a picture.

Blinded works the other way round. Type the name once and every mention is found on every page: in the text, in the lettering of screenshots and scans, and inside logos. Draw a box round a logo once and every other copy is found by what it looks like, at any size. Each word shows how many times it was found, which is also the quickest check: if a name should be on every page and the count says twelve of forty, you know where to look.

Turn on the detectors and email addresses, phone numbers, web addresses, street addresses and names in signature blocks are proposed without being typed. Turn on labels and each removed thing becomes a consistent code such as [P1], so the document still reads. Save a draft if you are interrupted, and nothing has to be redone.

Check the result before you send it

Open the redacted copy and try to get the content back. Select across a bar and paste somewhere. Search the file for each name you removed. Look at the file name and the document properties. If anything comes back, the redaction did not work.

Blinded rebuilds every page of the export from pixels, so there is nothing under a bar to select and no metadata carried over from the original, and it lets you rename the file as you save it. It is still worth doing the check: it takes a minute, and it is the only proof you have.

Nothing is uploaded

A redaction tool that uploads the file has sent the unredacted document somewhere before it has removed anything. Blinded runs in your browser tab: the PDF is opened, searched and rebuilt on your own machine, and the page is not permitted by the browser to send anything anywhere. Documents up to 20 pages are free.

How to do it, in four steps

Open the PDF in Blinded. It stays on your machine.
Type each name once and pick each logo once, instead of marking every mention.
Check the counts and the list of marks, and add a box of your own for anything missed.
Redact and export under a new name, then try to select and search the result.

Who it is for

Anyone who has to send a document with something taken out of it: paralegals and lawyers, HR teams, analysts, researchers, and anyone redacting for the first time.

Open the tool

Questions

Can I just draw a black box over the text?

Not in an ordinary PDF viewer. A drawn box is an annotation on top of the page, and the text underneath can still be selected, copied and searched. The content has to be removed, not covered.

Does printing to PDF remove the covered text?

Often not. Printing to PDF usually keeps the text layer, so the words under the box can survive. Check the result by selecting and searching it.

How long does it take?

Usually a few seconds a page, so a few minutes for a long document, with progress shown as it goes. Scanned pages take longer because they are read first. Most of the time left is your own check.

Is it free?

Documents up to 20 pages are free. Longer ones need a license to export the finished file.

Also

Back home