How to use the Redact PDF tool
- Choose the PDF. PaperKit reads the text of every page so it can search it.
- Tick the presets you need, such as Email addresses and Card-like numbers, add any name or number of your own, and press Find and mark.
- Check each page. Turn on Draw mode and drag across anything else you want hidden, including signatures, photos or a scanned page. Select a box to resize or delete it.
- Choose the resolution for redacted pages. 150 DPI is a good default.
- Press Redact and download PDF. Read the verification result, then save the file.
Worked example: a two-page bank statement
A 2-page statement shows [email protected] on page 1 and the card number 4111 1111 1111 1111 on page 2. With the Email addresses and Card-like numbers presets ticked, Find and mark reports 2 matches on 2 pages. The order number 12345 is not marked, because it is too short and fails the card checksum. At 150 DPI each US Letter page (612 by 792 points) is redrawn as an image of 1,275 by 1,650 pixels with the boxes burned in. The check afterwards reads the text of both pages of the new file and finds 0 of the 2 marked items. A third page with no marks would have been copied unchanged.
Why black boxes are not redaction
Most PDF editors let you draw a black rectangle, and many people stop there. The rectangle is a new shape lying on top of the page, but the text under it is untouched. Anyone can select it and copy it, search for it, or delete the rectangle. Court filings and government reports have leaked names and figures exactly this way.
True redaction removes the content itself. PaperKit does this by drawing each marked page into a picture, painting the boxes into that picture and replacing the page with it. The new page contains no text at all, only pixels, so there is nothing left to copy. The trade-off is that text on a redacted page can no longer be selected, which is why pages without marks are left exactly as they were.
Getting reliable results
Search only finds real text. A scanned page is already a picture, so draw boxes on it instead; it will be burned in the same way. Presets catch common formats but cannot know every way a number can be written, so always look at each page before you redact. If the check reports that a marked item is still present, it names the page: usually the same detail also appears on a page you did not mark. The new file also leaves out the original metadata, bookmarks, attachments and form data, which can hold the same private details.
Your files stay on your device
Redaction is only useful if the unredacted original stays private. This tool reads, rasterises and rebuilds the PDF entirely inside your browser, using pdf.js and pdf-lib running on your device. The original and the redacted copy are never uploaded, so there is no server copy that could leak.
Frequently asked questions
Can the redacted text be recovered?
Not from the redacted pages. Each one is replaced by an image with the boxes painted in, so the original text is no longer part of the file. The automatic check confirms it.
Why can I no longer select text on some pages?
Those are the pages you redacted; they are now images. Pages without any marks keep their selectable text.
Which numbers does the card preset find?
Runs of 13 to 19 digits, optionally in groups, that pass the Luhn checksum used by payment cards. That avoids marking most order or invoice numbers.
Does it work on scanned PDFs?
Yes, by drawing boxes. A scan has no searchable text, so presets find nothing, but drawn boxes are burned in the same way.
Will the file get bigger?
Usually a little, because each redacted page becomes an image. Use 100 or 150 DPI and JPEG to keep it small, or compress it afterwards.
Is it safe for legal documents?
It removes the marked content and checks the result, but you are responsible for marking everything that must be hidden. Review every page before you share the file.