How to Redact a PDF Properly (and Why Black Boxes Fail)

To redact a PDF properly, use a tool that removes the hidden text from the file, not one that just draws a black box over it, then test the result by searching for the redacted words and copying the blacked-out area. A black rectangle on top of text hides it from the eye, but the words are still in the file for anyone who copies and pastes.

This is not a theoretical risk. Courts, government agencies and law firms have published "redacted" documents whose secrets were one copy-and-paste away. This guide explains why that happens, how to redact in a way that holds up, how to check your own work, and which hidden data to clean out at the same time.

Key takeaways

• The NSA's guidance on redaction says "The most common mistake is covering text with black" (NSA, Redacting with Confidence).
• True redaction removes the content from the file. Siliph's Redact PDF rebuilds each affected page as an image with the black boxes burned in, so the hidden words cannot be copied or searched.
• Always test: search the redacted PDF for each hidden term and try selecting text under the boxes.
• Redaction removes visible content only. Clear document properties and other hidden data separately, for example with Remove Metadata.
• Scanned pages have no text layer, so run OCR before redacting by search term.

Why black boxes fail

A PDF page is built in layers. Text sits in one place, images in another, and annotations such as highlights, comments and drawn rectangles float on top. When you draw a black box with a highlighter or shape tool, you add a layer over the text. The text underneath is untouched. Select it, copy it, paste it into a notes app, and it comes straight out.

The NSA's report on the problem, written for agencies converting Word documents to PDF, is blunt: "The way to avoid exposure is to ensure that sensitive information is not just visually hidden or made illegible, but is actually removed from the original document" (NSA, 2005).

Two real failures

• A 2019 court filing. Paul Manafort's lawyers filed a document with sections covered by black bars. As the Columbia Journalism Review reported, a Guardian reporter found that all he had to do was "copy and paste some of the redacted text... into a new document, and the redactions magically disappeared" (Columbia Journalism Review, 10 January 2019).
• A 2009 security manual. A US Transportation Security Administration screening document was exposed "because of inadequate redaction procedures" and could be read by "anyone who copied the document and pasted it into another format, such as Microsoft Word or Windows Notepad" (Nextgov, 9 December 2009).

Both were made by professionals with access to proper tools. The mistake was the method, not the software.

What proper redaction does

Proper redaction has two parts: the content is removed, and the area is marked so readers can see something was taken out. Adobe describes it this way for Acrobat Pro: "redacted content... is replaced with colored boxes, ensuring that sensitive information cannot be recovered or viewed", and "Redaction is permanent and cannot be undone once applied and saved" (Adobe).

Adobe adds an important limit: "Redaction removes only visible text and graphics. Hidden data, such as metadata or embedded content, must be sanitized separately." The NSA report found the same trap in Word-to-PDF workflows, noting that "merely converting an MS Word document to PDF does not remove all metadata automatically."

How to redact a PDF with Siliph

Siliph's Redact PDF works by search term. It finds every match of the words, names or numbers you type, blacks them out, and rebuilds each page that had a match as a 200 DPI image with the boxes burned in. That deletes the text, links and comments on those pages. Pages with no match are copied unchanged, so they stay searchable.

1. Open Redact PDF and upload the PDF (one file per task, up to 200 MB).
2. Type a term in Text to redact. Click Add another term for each extra item: names, account numbers, addresses.
3. Add each format you expect separately. Capitals and spaces are ignored, so "1234 5678" also matches "12345678", but other characters such as dashes must match, so add "1234-5678" as its own term if it might appear that way.
4. Run the task and download the file, saved with -redacted in its name.
5. Test it, using the checklist below.

Two behaviours to plan for. Bookmarks and document properties are not carried into the redacted file, which removes a common leak. And because affected pages become images, their remaining text is no longer searchable; run OCR PDF on the result if you need it searchable again.

Scanned documents

A scan has no text layer, so a search term finds nothing. Run OCR PDF first to create searchable text, then redact the OCR copy. Our guide to making a PDF searchable covers getting clean OCR.

Signatures, photos and stamps

Redact PDF matches text only. To hide an image such as a signature or a photo on an ID card, cover it with a filled rectangle in PDF Editor, then turn that page into a flat image so the original picture is no longer underneath: export the pages with PDF to JPG at 200 DPI or more, and rebuild the PDF with JPG to PDF. A rectangle alone, even after flattening, can leave the original image in the file.

Test every redaction before you share it

Spend two minutes on this every time:

1. Search: press Ctrl+F (Cmd+F on a Mac) and search for each redacted term. Nothing should be found.
2. Select and copy: drag across a black box, copy, and paste into a plain text editor. Nothing from underneath should appear.
3. Check the properties: open the document properties and look at the title, author and subject. A file called "Settlement with J. Smith" gives away what the body hides.
4. Look for leftovers: check headers, footers, bookmarks, comments and attachments for the same names or numbers.
5. Check the file name: "smith-divorce-final.pdf" undoes careful redaction in the body.

If any test fails, start again from the original document, not from the failed copy.

Clean out hidden data too

Visible text is only part of a PDF. Documents can also carry an author name, the software used, creation dates, earlier titles and, in Word conversions, sometimes a copy of the source. The NSA report warns that Word's "Attach source file to Adobe PDF" option "inserts a copy of the original Word document inside the output file, almost certainly not what was intended."

Siliph's Remove Metadata clears the title, author, subject, keywords, creator and producer, resets the dates and deletes the hidden XMP data block. It runs in your browser, so the file does not leave your device for that step. Run it after redaction, then add a password with Protect PDF if the file still holds private details.

What US federal court filings must redact

If you file in a US federal court, Federal Rule of Civil Procedure 5.2 limits personal details in filings to "the last four digits of the social-security number and taxpayer-identification number", "the year of the individual's birth", "the minor's initials" and "the last four digits of the financial-account number" (FRCP 5.2). Other courts and countries have their own rules, so check the one that applies to you.

Other redaction tools

• Adobe Acrobat Pro: full redaction with mark-then-apply, plus a Sanitize document option for hidden data. It is a paid product (Adobe).
• Print to image: printing a document to paper and scanning it back removes hidden text, but you lose search, and it is easy to forget a page.
• Not suitable: highlighter tools, shape tools, and black text on a black background. All of them leave the text in place.

Redact PDF runs on Siliph's servers. The file travels over an encrypted upload, the search and redaction run automatically with no one viewing your pages, and both the upload and the result are deleted within 2 hours.

Frequently asked questions

Can redacted text in a PDF be recovered?

If the text was only covered by a black box or highlight, yes, by copying and pasting it. If a proper redaction tool removed the text from the file, there is nothing left to recover. Test by searching for the hidden words.

How do I know if a PDF was redacted properly?

Search the PDF for the redacted words, then select the black areas and paste them into a text editor. If either test reveals anything, the redaction failed. Also check the document properties for revealing titles or author names.

Can I redact a scanned PDF?

Yes, but text-based redaction needs searchable text. Run OCR on the scan first, redact the OCR copy by search term, then check that every instance was caught, since OCR can misread names and numbers.

Does redaction remove metadata?

Not necessarily. Redaction removes visible content. Siliph's Redact PDF drops bookmarks and document properties from the output, but run a metadata cleaner and check the file name too.

Is redacting a PDF free on Siliph?

Yes. Redact PDF is free with no account. Files are processed automatically over an encrypted connection and deleted from Siliph's servers within 2 hours.