Your redacted text is still extractable. Here is why
Every few months a redacted document is published, somebody selects the text under the black bars, and the whole thing is in the open by lunchtime. It is almost never carelessness. It is a tool doing exactly what it said it would.
Open the tool Free, in your browser. Nothing is uploaded.
After
What a black rectangle actually is
A PDF is a list of drawing instructions. "Put this text here in this font" is one instruction; "fill this rectangle with black" is another. Drawing the rectangle adds an instruction -- it does not remove the one before it. Both are still in the file, and the order they are drawn in is the only reason you cannot see the first.
So selecting across the bar copies the text. So does any PDF library, in about four lines. The same is true of a white rectangle, a highlight set to opaque, an image pasted over the area, and a shape added in a word processor before printing to PDF.
The other two ways it leaks
Even when the text is genuinely removed, the thumbnail embedded in the file may still show the original page, and the document metadata -- title, author, the original filename -- routinely carries the name that was removed from the body.
And a cropped image is not a cut image: cropping in most tools stores the whole picture with a smaller window onto it.
How to check the file you already have
Open the exported PDF, select across a redacted area, and paste somewhere. If anything comes out, the text is still in there. Search the file for a name you removed. Look at the document properties.
That is a two-minute test, and it is worth doing on whatever you use now before you need the answer.
What removal actually requires
The page has to be rebuilt. Blinded renders each page, draws the bars into the pixels, and writes a new file out of the result: there is no text layer under the bar because there is no text layer at all, until you ask for a fresh one to be read back from the visible page.
That costs the original’s selectable text, which is a real trade. It buys a document where what is covered is gone, which is the only property that matters once the file is out of your hands.
How to do it, in four steps
- Open the document that needs redacting -- in the tab; nothing is uploaded.
- Mark what has to go, by typing it, by picking it out, or by drawing on the page.
- Press Redact and check the marks against the pages.
- Export. The pages are rebuilt, so select-and-paste under a bar returns nothing, because there is nothing there.
Who it is for
Anybody who has been handed a redacted PDF and wondered whether it is really redacted, and anybody about to send one.
Questions
Does Acrobat’s redaction tool work properly?
Its dedicated redaction tool does remove content when the redaction is applied. The failures happen when a shape, highlight or image is drawn over the text instead, which looks identical on screen and removes nothing.
Can I recover text from a badly redacted PDF?
Usually, yes -- selecting and copying across the bar is often enough. That is the point: if you can do it in a minute, so can the recipient.
Is the exported file still searchable?
Not by default, since the page is a picture. There is an option to read the redacted pages back so the remaining text can be found and copied -- which reads only what is visible, so the removed words cannot come back.
What about the metadata and the thumbnail?
The exported file is written fresh from the rebuilt pages rather than edited, so it does not inherit the original’s embedded thumbnails. You can also rename the file on export, since a filename is often the last place a name survives.