Really redact a PDF
Mark what has to go and the text disappears from the file, instead of being covered by a rectangle. It runs in your browser: the document is never uploaded.
Covering something with a black rectangle is not redaction. When you draw a box over an ID number in a PDF viewer, the number is still underneath: it can be copied with the mouse, a search engine finds it, and it shows up in full when the file is opened in another program. It has happened in courts, in ministries and in newspapers, and it is always found out the same way: somebody selects the text and pastes it. The same goes for the white highlighter and for pasting an image on top.
Here the text is removed from the content stream, which is where a PDF really keeps what each page says. The marked glyphs disappear and the ones left are repositioned so the line does not move. If there is an image under the box, its pixels are erased and the image is packed again. When it finishes, the file is read back with a different engine to check that nothing copyable is left there, and you are told the result.
FAQ
How do I know the text is really gone?
Because it is checked, not promised: when it finishes, the PDF you are taking away is read again and text is searched for under every box. If anything were left, you would be told.
Can I find the data instead of marking it by hand?
Yes. The find personal data button goes through the whole document for ID numbers, bank accounts, emails and phone numbers and marks them. Each one is validated first, so it does not fire on any random number.
Does it work with scanned documents?
Yes, marking by hand: in a scan the page is an image and what gets erased are the pixels. What is not there yet is automatic search inside a scan, which needs OCR.