How to Redact a PDF Properly

Redaction failures are among the most repeated document-security mistakes. Courts, government agencies, and companies have all published PDFs where 'redacted' text could be recovered by anyone who selected the black box and pressed copy. The failure is always the same: the text was covered, not removed.

A PDF is not a flat picture. It is a container of layered objects: text streams, fonts, images, and annotations. Drawing a black rectangle adds a new object on top of the text object. The text is still there, still selectable, still searchable, and still extractable by any PDF library.

How redactions fail

The classic failure is the drawn shape: a black rectangle or highlight-set-to-black placed over sensitive text with an annotation tool. Copy-paste defeats it instantly. A subtler failure is deleting the visible text but leaving it in places you cannot see: PDF metadata, embedded attachments, form field values, comments, bookmarks, or an invisible OCR text layer sitting behind a scanned page.

A third failure is redacting a derivative while the original leaks. If the PDF's metadata records the source document's title or author, or if an earlier revision is embedded, your careful page-level work can be undone by a single properties panel.

What proper redaction requires

True redaction removes the sensitive content from the file's internal structure so that no tool, however clever, can get it back. SecretPNG's redaction tool does this by flattening redacted pages to pixels: the page is rendered to an image with the redacted regions removed, and the output page contains only those pixels, with no text objects underneath. The tool then re-extracts text from the output file and checks that none of the redacted content survived, as a built-in validation step.

Flattening trades some file properties for certainty: flattened pages are no longer text-searchable and can be larger. For genuinely sensitive material, that is the right trade.

A redaction checklist

Before releasing a redacted document, walk through each step.

• Redact with a tool that removes content, never with drawing or highlighting tools.
• Clear document metadata: title, author, subject, keywords, and creation tool fields.
• Check for comments, form fields, attachments, and bookmarks that may repeat the sensitive content.
• Open the final file and try to select, search for, and copy the redacted material.
• If the stakes are high, extract all text with a separate tool and search the output.

Inspect the export, every time

Automated validation is a strong safety net, but no tool can know your document's context: a name you forgot to mark, an account number that appears on a second page, a detail visible in an embedded image. Always review the exported file page by page as the final step. Redaction is one of the few security tasks where a two-minute manual check is both feasible and decisive.

Also remember that redaction and encryption solve different problems. Redaction permanently removes content for recipients who should see the rest; encryption protects the whole document from everyone without the key. Sensitive documents often need both: redact what the recipient should not see, then encrypt the file for delivery.