Before you share, check what stays behind.

A black rectangle is not automatically a redaction. If the original text remains beneath it, someone may still be able to copy or search it. Treat redaction as a removal task that needs verification.

Identify every occurrence

Make a list of the words or phrases that need removal. Include variants: a name may appear with initials in one place and in full in another. Check headings, footers, appendices, and any repeated table entries.

PDF AI’s Redact PDF workflow searches selectable text. It is not a visual marking tool for arbitrary regions of a scan. If the sensitive information is embedded in an image, do not assume text matching will find it.

Create and inspect the redacted copy

Upload the PDF and enter the exact phrases, one per line. Run the tool, download the result, and open that downloaded file in a PDF viewer.

Search for each removed phrase. Try selecting and copying text around the redacted area into a plain-text editor. Inspect every expected match and confirm that nearby content still makes sense. A “completed” message only tells you the operation finished; it does not confirm that all sensitive information was found.

Look beyond the visible page

A PDF can include metadata, attachments, links, and other material. Sanitize PDF removes selected categories of embedded data, but sanitizing and redacting are different operations. Neither should be treated as a blanket guarantee that a complex file contains nothing private.

Keep the right version

Store the original in its approved location and name the sharing copy clearly. For high-sensitivity material, use a reviewed redaction process and an independent check before distribution. Never rely on cropping a page or drawing a box over words as proof of removal.