Redact the PDF. Check the copy.

Excise starts by proposing common private-data patterns. You decide which to add, then add any names, numbers, or phrases of your own.

• Automatically searches for six formats: Social Security numbers, card numbers, IBANs, email addresses, phone numbers, and dates. Date search is off by default.
• On Mac or iPad, select text on the page and mark this occurrence or every matching occurrence across the PDF.
• Add as many searches as the job needs in the left panel. Each rule shows how many page-text matches it found.
• Draw a rectangle around a signature, stamp, image, or mixed area that text search cannot name.
• Nothing is applied automatically. Proposals and matches move to Review.

Excise shows every match in context and every area you marked. On Mac and iPad, uncheck any occurrence you want to leave in the copy.

• See the surrounding words, page number, replacement, and any exemption code for each match.
• Marked areas show what they cover and whether the page will be converted to an image.
• On Mac and iPad, review one occurrence at a time; an unchecked match is deliberately left in place and keeps the result from passing.
• Press Redact only after the list is right. Excise writes a separate staged copy; your original stays unchanged.

Excise reopens the staged redacted copy and bases its result on what it reads back, not on what the run intended to do.

• Switch between Original and Redacted copy before you save.
• It re-extracts every page's text and checks the document structure for strings your rules still target.
• Fourteen supported hidden-data categories are cleared by default, including metadata, form fields, bookmarks, annotations, and embedded files.
• Excise keeps unaffected text live when it can do so safely. When it cannot, Excise can convert the page to an image to complete the redaction; the receipt tells you when it does.
• The result is explicit: Verified, Review required, or Not verified.

How Excise checks the saved copy →