How to Redact a PDF Properly (and Permanently)

Redaction means deleting content so that no method can bring it back. In a PDF, sensitive data can sit in several places at once: the visible text and graphics, a hidden text layer on scanned pages, embedded images and metadata inside the file.

The test is simple. After redaction, the words, numbers or images you removed must not exist anywhere in the file. Opening it in a text editor, searching, copying the page or reading the raw bytes must all come up empty. Anything less is hiding, not redaction, and hiding fails.

The most common mistake is treating redaction as a drawing task. A black rectangle, a black highlight, text colored to match the background or an image pasted on top all look convincing. The original text is still there. Select the page, copy it, and the hidden words paste into any text editor.

This happens in real cases. In January 2019, lawyers for Paul Manafort filed a US federal court document whose blacked-out text could be read by copying and pasting it. In December 2009, the US Transportation Security Administration posted its airport screening manual online with redactions that could be removed to show sensitive procedures. In both cases a box was drawn over the text instead of the text being deleted.

Proper redaction deletes content. It doesn't paint over it. A real redaction tool removes the characters, vector shapes and image pixels in the marked area, then draws a black bar over the empty space. The PDF standard (ISO 32000) defines a redaction annotation, but it only marks content for removal. The data stays in the file until you apply the redaction as a separate step.

The visible page is only part of the job. The same information can sit in metadata, comments, form field values, bookmarks, scripts, attachments, hidden layers, image EXIF data and the invisible OCR text layer on scanned pages. The OCR layer is the one people miss most. Any of these can leak what you thought you removed.

Use software with a real redaction feature that marks content and then deletes it. Editors that only let you draw shapes won't do. The workflow is the same in any proper tool. Mark what to remove, then apply the marks to delete it.

For scanned pages, or when you need full certainty, turn the redacted pages into plain images. This removes any leftover text layer, but the text on those pages can no longer be selected. PDFToolz Redact PDF works this way. It redraws each marked page as an image with the black boxes burned in, so the text and images underneath are gone.

Don't assume redaction worked because it looks right. Checking takes about a minute. If any of these checks shows content you removed, the file is not safe to share.

Once the visible page is clean, three hidden sources cause most leaks. The first is metadata. The document info dictionary and XMP packet can store the author, original title and keywords, sometimes including the exact text you redacted. The second is incremental saves. Each save adds new data without removing the old, so earlier versions stay in the file until you use Save As to write a fresh copy.

The third is the OCR layer, and it is the one people miss most. Scanned pages have an invisible, searchable text layer behind the image. A black box over a scanned name leaves the name searchable and copyable unless that text layer is removed too.