Redact a PDF

Why "redaction" usually is not

A PDF is a set of drawing instructions. Adding a black rectangle adds one more instruction: draw a black rectangle here. It does not remove the instruction that draws the words, which still sits in the file, in order, fully intact.

So the document looks censored and is not. Select the blacked-out area and the text highlights. Copy and paste it and it pastes. Run any extraction tool over it and it comes straight out. The black box is a picture of secrecy laid on top of the secret.

This is not an obscure failure. It has exposed the names of protected witnesses, unredacted settlement figures, the identities of intelligence sources and the terms of sealed agreements — each time in a document that somebody looked at, saw covered in black bars, and reasonably believed was safe.

What this does instead

Every page you mark is rendered to pixels at the resolution you choose. The black rectangles are painted onto that raster, and the original page — content stream, fonts, text objects and all — is thrown away and not copied into the output.

Afterwards the covered area is black pixels and nothing else. There is no layer beneath it, because the layer was not hidden, it was discarded. That is checkable in the obvious way: open the result and try to select the blacked-out region, or put it through the text extractor and see what comes out.

The cost, stated plainly

A flattened page is no longer text. It cannot be searched, selected, copied from, or read aloud by a screen reader. That is not an implementation shortcut — it is inherent. A page cannot simultaneously keep its text layer and have part of that layer destroyed, because they are the same thing.

What can be limited is the blast radius, and it is: only pages carrying a mark are flattened. A hundred-page disclosure bundle with two censored pages keeps ninety-eight of them fully searchable.

What redaction does not protect you from

It removes what you mark. It cannot remove what you did not think of, and the leaks that follow a technically correct redaction are almost always about context rather than pixels.

The same name appearing unredacted three pages later. A covered figure whose value is implied by a total elsewhere. A box whose width gives away the length of the word beneath it. A filename, an email subject line, or a covering letter that names what the document is about.

The document properties are handled here — title, author, subject and keywords are cleared on the way out, because censoring a body while leaving the case name in the title achieves very little. The rest is a reading job, and it is worth doing before you send.

Questions

Does this actually remove the text, or just cover it?

It removes it. The marked pages are rendered to images and the original pages are discarded, so the characters no longer exist anywhere in the file — you can confirm it by opening the result and trying to select the blacked-out area, or by running it through a text extractor. Drawing a black rectangle over live text, which is what most tools do, hides nothing.

Why do the redacted pages stop being selectable text?

Because that is the price of the text genuinely being gone. A page cannot both keep its text layer and have part of it destroyed — the two are the same thing. Pages you did not mark are copied across untouched and stay fully searchable, so a long document only loses the pages that needed censoring.

Is my document uploaded?

No, and consider what redaction is usually for: legal disclosure, medical records, financial statements, personal data in a subject access request. Those are the last documents that should be handed to an unknown server, and there is no need to — this runs entirely in your browser and makes no request at all.

Can redacted text ever be recovered?

Not from the pixels. Once a page is a picture with black rectangles painted on it, the covered area contains black pixels and nothing else — there is no layer beneath to recover. What can still leak is context: page numbers, a distinctive line length, or the same name appearing unredacted elsewhere in the document. Redaction removes what you mark, not what you forgot.

What about the document properties?

Cleared. The title, author, subject and keywords are removed from the output, because redacting a body while leaving "Smith v Jones — settlement draft 3" in the title deletes the answer and keeps the question.

What resolution should I use?

150 DPI is readable on screen and keeps the file small. Use 200 to 300 if the document will be printed or if it contains small print that has to stay legible. It only affects the pages you marked; everything else is untouched.

What BriskFile will not do to you

• Your files never leave your device

  Every conversion runs in your browser. Open DevTools, watch the Network tab, and you will see no upload — because there is not one.

• No account, no email, no watermark

  Nothing to sign up for and nothing stamped on your images. There is no step between choosing a file and getting it back.

• No limits at the download button

  No daily cap, no file counter, no "upgrade to download". If the tool starts a job, it finishes it.

• Checkable, not just claimed

  Open the Network tab and convert something — nothing goes out. Or load the page, disconnect, and watch it keep working: the tool is already on your machine, and your file never leaves it.