How to redact a PDF so the text is really gone

Why a black box is not redaction

Every few months a court filing, a government release or a company memo makes the news because someone drew black rectangles over names and the names were still there. The rectangle was a drawing on a layer above the page. The text underneath was untouched. Anyone could select and copy it, or delete the rectangle.

PDFs are built in layers. The page content stream holds the text and images. Annotations such as highlights, shapes and comments float on top. Most of the quick fixes people reach for live in that top layer: the rectangle tool in Preview or Acrobat Reader, a shape in Word before export, a highlighter set to black. None of them touch the content stream.

There are more hiding places. The document metadata keeps the author and title. Bookmarks can quote headings. Form fields keep their values. A scanned document that was run through OCR carries an invisible text layer that copies even though you cannot see it.

What true redaction does

A real redaction tool changes the page itself. For text, it finds the characters inside the marked area and deletes them from the content stream, then draws the black box so the layout does not collapse. For images, it removes the pixels in that area, not just covers them. The result is flattened so there is nothing to peel back.

A thorough tool also sanitizes: it clears the metadata, removes comments, bookmarks and attachments, and drops hidden layers. After a true redaction, selecting all the text on the page and pasting it into a text editor shows the surrounding text with a gap where the name was.

• Redaction removes content. It cannot be undone, so work on a copy.
• Redaction is final only once you save. Many tools show a preview, and the text is gone when the new file is written.
• Covering, highlighting and white-out are not redaction. A redaction mark that has not been applied is just a shape.

How to redact a PDF online for free

Without Acrobat Pro, a browser tool is the fastest route on any device. Redact PDF on DocGopher blacks out names, email addresses, phone numbers, IBANs and other text, and the text is deleted, not just covered. You can type the exact words to remove, or let it find every email address or phone number by pattern, so a 40-page statement with the same account number on each page is handled in one pass. The document metadata is removed as well. It is free, needs no sign-up and adds no watermark. Files up to 300 MB are processed in memory on DocGopher's own servers and deleted after 30 minutes.

A search-based tool only finds text that exists as text. If the PDF is a scan without a text layer, see the section on scanned documents below.

1. Open Redact PDF in your browser and drop in the PDF, or up to 50 files to treat the same way.
2. Type the terms to redact: a name, a Social Security number, an address. Add one per line.
3. Turn on the pattern options you need, such as all email addresses or all phone numbers, so you catch the ones you did not think of.
4. Run the redaction and download the new file.
5. Open the result and run the checks in the section below before you send it.

How to redact a PDF with Preview on a Mac

Since macOS Big Sur, Preview has a real Redact tool, and it removes the content rather than covering it.

Two cautions. Preview's redaction removes what is under the box, but it does not clear document metadata, so check File, then Tools, then Show Inspector for the author and title afterwards, and recreate the PDF if you need those gone. And on older macOS versions, the only black rectangle available is the shape tool, which is a cover, not a redaction.

1. Open a copy of the PDF in Preview.
2. Click the Markup toolbar button, then choose the Redact tool (the icon looks like a black marker). Preview warns you that redaction permanently removes the content; click OK.
3. Drag across the text or image you want to remove. It turns black at once.
4. Repeat for every item on every page. Use Cmd+F to find other occurrences of the same name.
5. Choose File, then Save (or Export). The redactions are applied when the file is written.

How to redact a PDF in Adobe Acrobat

Acrobat Pro has had redaction for a long time, and it is the standard in law offices. The free Acrobat Reader cannot redact; its rectangle tool is just a shape.

1. Open the PDF in Acrobat Pro and choose Redact from the tools list.
2. Use Mark for Redaction to drag over text or images, or Find Text to search for a word or a pattern such as phone numbers and mark every match at once.
3. Set the appearance if you want a label such as a reason code instead of a plain black box.
4. Click Apply. Acrobat asks you to confirm because the change is permanent.
5. When prompted, also choose Sanitize document (or Remove Hidden Information) to clear metadata, comments, bookmarks and hidden layers. Save under a new name.

How to check that the redaction worked

Never trust the black box. Spend two minutes on these tests with the saved file, not the one still open in your editor.

If anything fails, go back to the original copy and redact again with a tool that actually removes content. Do not try to fix the already redacted file by stacking another box on it.

1. Copy-paste test: open the redacted PDF in any viewer, press Cmd+A or Ctrl+A to select everything on a page, copy, and paste into a plain text editor. The redacted words must not appear.
2. Search test: press Cmd+F or Ctrl+F in the PDF viewer and search for the redacted name, a few digits of the account number, and the email domain. Zero results is the only acceptable answer.
3. Conversion test: run the file through PDF to Word or PDF to Text and search the result. Conversion pulls out every character the file still contains, including text hidden under images.
4. Metadata test: open the document properties (File, then Properties in Acrobat or Chrome; Show Inspector in Preview) and check Title, Author, Subject and Keywords.
5. Annotation test: if your viewer has a comments or annotations panel, make sure it is empty. A redaction that still shows up as an annotation was never applied.

Redacting scanned PDFs, and what OCR changes

A scanned PDF is a picture of a page. There is no text to search, so a search-based redaction finds nothing, and a copy-paste test passes even on an unredacted scan, which can be misleading. To redact a pure scan, you need a tool that removes pixels in a drawn area, such as Preview's Redact tool or Acrobat's Mark for Redaction.

Many scans are not pure, though. If a scanner, a copier or an earlier step ran OCR, there is an invisible text layer sitting on top of the image. That layer copies and searches like normal text. A cover drawn over the image hides the pixels but leaves the hidden words in place. A proper redaction tool removes both. This is why the conversion test above matters: it catches the hidden layer.

If you want a scan to be both searchable and redacted, the safe order is to redact the image first, then run OCR on the redacted file. OCR can only read what is visible, so the redacted areas stay blank in the new text layer.

Do not forget the metadata and the file name

After the page content is clean, three things still leak information. The metadata fields can hold the author's name, the company, the software and sometimes the original title, which might be the client's name. The file name itself often carries a name or a case number, so rename the redacted copy before it leaves your computer. And the box itself can leak: one the exact width of a 12-letter surname says more than one covering the whole line, so some tools let you redact a full line instead of only the matched text.

Finally, think about where the original lives. A redacted PDF sent by email is pointless if the unredacted version sits in a shared folder the same recipients can open. If the file must also stay confidential within a group, add a password with Protect PDF and send the password separately.