Secure & archive · guide
What Redact PDF does
Redact PDF rebuilds each page you mark as an image with the marked areas painted solid black, instead of drawing a box that sits on top of the text. Before you get the file, every rebuilt page is checked for text that could still be extracted, and if any is found you get an error instead of a file. Pages you do not mark are copied across rather than redrawn, so check them in the result: in a PDF your reader shows as secured or encrypted, even one that opens without a password, they come out blank, and content on a hidden layer becomes visible. The document never leaves your browser, which matters most for the very files you would redact.
How to redact a PDF properly
- Choose one PDF Select the document to redact. It opens in this browser tab and is not uploaded; every page is drawn so you can mark it, which takes a moment on a long file.
- Mark private content Drag across the page to draw a box over every area that must go, including details repeated on other pages and in headers or footers. Click a box to remove it, and review every mark before saving.
- Redact and save Select Redact and save. If any text could still be extracted from a redacted page, no file is produced; otherwise save the copy and test it by trying to select text in the marked areas.
Worked example
Remove an account number from a statement copy before sharing.
- Starting point
- Choose the statement and identify every place the number occurs, including images and repeated headers.
- What to do
- Draw a redaction area for every occurrence using the manual page controls, review all marks, export and inspect the tool’s result notes.
- Result
- Expect a new redacted copy. Reopen it, search for the removed value and visually inspect the affected areas and any additional occurrences.
If the result is not right
For a scanned number that search misses, mark the visible area and inspect the exported page. Keep the original securely and do not share a result until all intended occurrences are checked.
Limits worth knowing
- A black rectangle in an ordinary editor is not redaction. No single search proves removal of every duplicate, image or alternate representation.
- This workflow stops on interactive forms rather than risk losing their fields. Use Flatten first and check every visible answer in that saved copy, then open the flattened PDF here. Keep the editable original; protected or XFA forms may need a compatible form application first.
Questions and answers
Could someone lift my black boxes off and read what is underneath?
A box drawn in an ordinary PDF editor can be lifted, because the text is still there beneath it. Here each page you mark is rebuilt as a single image with the boxes painted into its pixels, so on that page there is no separate box to move and no text layer under it. The tool then re-opens the result and checks every rebuilt page for extractable text; if any turns up, you get an error instead of a file.
My bank statement repeats the account number on every page. Do I have to mark each one?
Yes. There is no search or apply-to-every-page option, so each occurrence needs its own box, including headers, footers and reference lines that repeat the number. Search the original in your usual PDF reader first to find every place it appears. Each page you mark becomes an image, so a statement marked on every page can no longer be searched at all.
Does it work on a scanned document or a photo-only PDF?
Yes. You mark what you can see, and a marked page is rebuilt as an image whether it started as text or as a scan; any hidden OCR text on that page is checked along with everything else. Pages you leave unmarked keep their OCR layer, so make sure every page that shows the detail is marked.
Why is the redacted file bigger, and why can't I select text on marked pages?
Each marked page is redrawn as a picture at 150 DPI, or a little less on very large pages, which usually takes more space than text and can look slightly softer when you zoom in, and a picture has no selectable text. Unmarked pages are copied rather than redrawn, so they keep their text, apart from the cases listed under what redaction does not remove. If the redacted pages need to be searchable again, run OCR PDF on the result; the black areas stay blank.
What does redaction not remove?
Anything outside your boxes, on marked and unmarked pages alike, so check every page before sharing. Unmarked pages need a look in particular: in a PDF your reader shows as secured or encrypted, even one that opens without a password, they come out blank, and content on a hidden layer becomes visible. It also does not look for hidden copies: an unmarked page is copied with everything it refers to, so if it links to a marked page or shares its images, the original can come along inside the file, where a viewer shows nothing and the check does not look. A small box in an empty margin of every page rebuilds every page as an image and copies nothing from the original, at the cost of all selectable text. The new file does not take the original's title, author, bookmarks or document-level attachments, but a file attached to a comment on a page you did not mark is copied with that page. It is named after your original with -redacted added, and a file name such as a client's name travels with it: use Rename before saving if it gives something away.
Is it safe to redact a confidential file in the browser?
The pages are drawn, rebuilt and checked inside this tab. The page's code and the PDF engine's font files come from filozy.com, but the document is read from your device and never sent; the receipt under the result shows 0 document bytes sent. Closing the tab clears the open file from memory.