Skip to content
Guides6 min read

How to redact a PDF properly: a step-by-step guide

A black box does not remove text from a PDF. Learn the safe routes: mark, apply, remove hidden data, or flatten to images. Then verify the file.

By Alexis de ONYRI

To redact a PDF properly, the text must be removed from the file, not just covered. Use a real redaction tool: mark the passages, apply the redaction, remove hidden information, and save under a new name. Or turn every page into an image with opaque bars. Then test the result before you send it.

Why is a black box or a highlighter not redaction?

A shape, a highlight or a comment sits on top of the page. The words stay in the file underneath. A reader can select them, copy them, search for them, or simply delete the shape.

The UK regulator, the ICO, shows this in its guide How to disclose information safely (version 1.2, May 2018). It hid text with the black highlighter in Word and exported the file to PDF. A copy and paste into a text editor revealed it. For the ICO, redaction must irreversibly remove the information from the released copy. The Lower Saxony data protection authority (Germany, July 2025) adds that white text on white fails too, and that software can sometimes reverse blur or pixelation.

How do you redact with a dedicated redaction tool?

Dedicated tools work in two stages: you mark, then you apply. Only the second stage removes the content. Adobe's Acrobat tutorial describes this flow. It says the tool is available only in Acrobat Pro, Premium or Studio (as of September 2026). Other PDF editors use their own menu names.

  1. 1Open the redaction tool. In Acrobat it is called Redact a PDF.
  2. 2Mark every passage: select text and images, mark whole pages, or search for a word, a phrase or a pattern.
  3. 3Review the marks page by page. A mark is only a preview. Check headers, footers, footnotes, tables, signatures and stamps.
  4. 4Select Apply. This is the step people forget.
  5. 5Select Continue to sanitize and remove hidden information, such as comments, metadata or hidden layers. Adobe says the redactions and hidden information are then permanently removed and saved to a new file.

Take Jane Example. She must send a contract without her home address. A search for her surname finds every Example in the text. It misses the line with only her initials. It misses her street in a scanned annex, because a scan is a picture. Search helps. It does not replace reading.

Can you redact a PDF with Preview on a Mac?

Yes. Preview has a redaction tool, and Apple documents it. In the Preview User Guide it is called Redaction Selection. Apple's description is short: select text to remove it from view. You can change the redaction as you edit, but once you close the document, it becomes permanent.

Three caveats. First, Shapes, Sketch, Draw and Highlight Selection in the same toolbar are annotations, not redaction. Second, Apple notes that if the PDF has a table of contents, entries linking to redacted text are not removed. Third, Apple speaks of removing text from view. Its page does not say what happens inside the file. It mentions neither metadata nor scanned pages. So close the file, reopen it, and run the checks below.

Can you redact in a browser, or by printing to PDF?

Browser PDF viewers are made for reading and light notes. Microsoft's documentation for the PDF reader in Edge, for example, lists inking, highlights and text notes, but no redaction feature (page checked in September 2026). Ink over a name is only an annotation.

Printing to PDF feels like a fresh start. It is not reliable. In its Word example, the ICO writes that even a Save as PDF or Print as PDF function is unlikely to provide effective redaction: the PDF can keep the highlighter mark along with the text. Apple's guide says File, Print, Save as PDF flattens annotations so that they cannot be edited. It does not say the words under them are gone.

Online redaction sites raise another question: where does the file go? The Lower Saxony authority advises against online tools that make you upload the original to the cloud. Read how a service handles files before you use it.

Is turning the pages into images a safe route?

Flattening means each page becomes a picture, with opaque bars painted into the pixels. No text layer is left, so nothing can be copied from under the bars. The ICO notes that redaction software generally works this way. LibreOffice documents a similar route for its own documents. According to its help, Tools, Redact opens a copy as a drawing, and the Redacted Export writes it as pixels into a PDF. This prevents any attempt to restore or copy the original contents.

The price is usability. The ICO points out that the result is no longer machine-readable: the text cannot be extracted or processed further. Nobody can search the file, and a screen reader finds no text to read aloud. If the recipient needs a searchable or accessible document, a dedicated tool that removes only the marked passages fits better.

Some newer tools add automatic detection to the flattening route. ONYRI Sanitize is one example: detection and masking run inside the browser, the file is not uploaded, and the export is always flattened. No detector finds everything, so you still review each proposed item before you download.

How do you check that the redaction worked?

  1. 1Take the final file, the one you will send. Close it and open it again, ideally in a different PDF viewer.
  2. 2Select all, copy, and paste into a plain text editor. In a flattened file, nothing should be selectable.
  3. 3Search for each removed value and for parts of it: the surname alone, the last four digits of an account number.
  4. 4Extract the text with a separate tool, for example the pdftotext command-line utility from Poppler, and search the output.
  5. 5Open the document properties: title, author, subject, keywords. Check the file name too: contract-jane-example.pdf gives the name away.
  6. 6Open the side panels: table of contents or bookmarks, comments, attachments, layers.
  7. 7Zoom in on each bar. It must be fully opaque and cover the whole value.
  8. 8Read the page as a stranger would. A job title plus a date can still point to the person.

Redaction serves data minimisation, as the Lower Saxony authority notes: under Article 5(1)(c) of the GDPR (EU), personal data must be limited to what is necessary. But hiding names only reduces exposure. Other details may still identify a person, so the document is not automatically anonymous. When the stakes are high, ask a second person to run the checks.

Frequently asked questions

Can someone remove a black box from a PDF?
Yes, if the box is a shape or an annotation. It can be deleted or moved, and the text under it can be copied. If a real redaction tool was applied, or if the page was flattened to an image, nothing should be left under the bar. Test it to be sure.
Does redaction also remove metadata?
Not automatically. In Acrobat, removing hidden information is a separate step offered after you select Apply. In other tools, open the document properties yourself and check the title, author and keywords. Check the file name as well.
Is a redacted PDF anonymous under the GDPR?
Not necessarily (EU). Recital 26 says the GDPR does not apply to information that does not relate to an identifiable person. But a document with masked names can still identify someone through context: a role, a date, a place. Redaction reduces the risk. Whether the result is anonymous must be judged case by case.

Sources & references

On this siteAnonymize a PDF in your browser

Mask a document without uploading it

ONYRI Sanitize finds names, identifiers, bank details and secrets in a PDF, a Word file or a scan, and masks them in your browser. You check the preview, then download a flattened copy.

Read next