How to redact personal data in a PDF

Why a black box isn’t enough

A black rectangle drawn with an annotation tool sits on top of the text, but the text is still in the file: it can be selected, copied, searched and often uncovered. Whoever receives the PDF only has to select the covered area and copy it to read what’s there.

The same goes for “Flatten PDF”, which only moves annotations into the page content: what’s underneath is still there. Only real redaction removes the text, images and drawings that were underneath.

Redacting a PDF, step by step

1. Open Redact PDF and drag the PDF into the workspace. You can also choose several PDFs to search for the same words in all of them, but then only “Search” is available and everything found is redacted, with no list where you can uncheck what you want to keep.
2. Choose how to mark what you want to cover: “Areas” (you drag on the page over a signature, a photo or a paragraph) or “Search” (you type words or pick patterns).
3. In “Search”, type names and phrases under “Words or phrases to redact” and pick from the “Quick patterns” (emails, phone numbers, ID numbers, IBANs, payment cards). You see every match and can uncheck the ones you want to keep.
4. Click “Redact PDF” and download the result. The original stays as it was.

What’s really removed from the file

Each page with redactions becomes an image with the black boxes painted in: the text, images and drawings underneath no longer exist, and neither does anything outside the visible area of the page. The metadata (including the metadata inside photos), digital signatures, scripts and the structure for screen readers are removed as well, along with comments and attachments unless you turn those options off.

MovaDoc writes the whole file again, so older versions that some programs keep inside the PDF disappear too. At the end, MovaDoc reads the result again and, if what you searched for or covered is still there, doesn’t hand it over.

In exchange, the text on redacted pages can no longer be selected, copied or searched, and the file may get bigger. Pages without redactions keep the text and appearance they had.

Scanned PDFs

In a scan, the pages are images with no text, so search finds nothing: use “Areas” and draw the boxes by hand. If you first recognize the text with OCR PDF, search works, but recognition can misread a letter or a digit, and search won’t find what was misread. Always check the pages with your own eyes.

Don’t forget the rest

• The metadata: the title, the author and the program that created the PDF. “Redact PDF” removes it from the result; if you only want to clear it, use Edit PDF metadata, but “Clear all” doesn’t remove the data inside photos (such as the location) or the names of whoever wrote the comments.
• The file name: the redacted PDF keeps the original’s name (with -redacted at the end), and “contract-mary-smith.pdf” already gives away what you wanted to hide. Rename it before you share it.
• Anything that identifies a person, not just their name: addresses, license plates, customer or case numbers, signatures, photos and logos.
• The rest of the text: covering a name doesn’t protect someone if the document describes them in another way.

What to check before you share

• Open the new PDF and go through every page: is everything that should be covered actually covered?
• Try to select and copy the text in the covered areas, and search (Ctrl+F; on a Mac, ⌘F) for one of the redacted words: nothing should come up.
• Look at the document properties: the title and author shouldn’t say anything you don’t want to share.
• Keep the original somewhere safe: what’s removed from the file doesn’t come back, and the original is the only complete copy.