Sensitive Data Redaction
Find email addresses, phone numbers, ID numbers, bank card numbers and your own words in a PDF, look over what was found, and remove what you choose for good.
About This Tool
Sensitive Data Redaction looks through a PDF for email addresses, phone numbers, ID and passport numbers, bank card numbers and IBANs, and for any words you type. It lists each thing it found, marked in the line it stands in and with its page, and removes the ones you leave ticked. What is blacked out is the address, number or word itself, not the line around it.
Drawing a black box over text is not redaction: the words underneath can still be selected, copied and searched. This tool redraws every page that has a redaction as a picture with the boxes painted in, so the text under them no longer exists in the file. It then reads the result back and refuses to hand it over if any text can still be extracted from those pages.
Pages without a redaction are left exactly as they were. Only the text printed on the pages is searched: comments, what is typed in form fields, bookmark titles, document properties and attached files are not. Arabic is read in the order it is written, and a number is found whether its digits are Western (0123) or Arabic (٠١٢٣). Everything happens in your browser.
How to Use
Upload your PDF
Choose the document that contains personal or confidential details.
Choose what to look for
Tick the kinds of data to look for, and type any names or other words of your own under "Other words or phrases", separated by commas. Then click "Find".
Look over what was found
Each find is listed under its kind, marked in its line. Untick anything that should stay: an order number of ten digits looks like a phone number. If something you expected is not in the list, type it as a word and click "Find" again.
Redact and download
Click the button under the list, which says how many places it will redact, then download the file and check it before you share it. To remove more, tick other kinds or type other words and find again: each round adds to the same file.
Use Cases
Commercial agreements sharing
Black out the phone numbers, email addresses and account numbers in an agreement before it goes to someone outside.
Resume database anonymization
Take the contact details out of a CV. Type the applicant's name and address as words of your own: a name has no look to be known by.
Financial statement distribution
Hide card numbers and IBANs in a statement, and any other account numbers or names that you type in.
Frequently Asked Questions
Can the removed text be copied or recovered?
No. A page with a redaction is replaced by an image of itself with the boxes painted in, so there is no text left under them to copy, search or extract.
What changes on a redacted page?
The whole page becomes an image: none of its text can be selected or searched any more, and links and form fields on it stop working. Pages with nothing to redact are untouched.
What does each kind find?
"Email addresses": anything written like [email protected]. "Phone numbers": numbers in the usual groupings, with the code of a country (+966 55 123 4567) or without one (055 123 4567, (555) 123-4567, 4471 2233), and ten to fifteen digits in a row. "ID and passport numbers": identity numbers of eighteen digits, numbers written like 123-45-6789, Emirates ID numbers, numbers of ten digits that begin with 1 or 2, and passport numbers of one or two letters and seven or eight digits. "Bank card numbers and IBANs": card numbers written in fours or in a row, and IBANs, each only if its check digits are right, so that a number which merely looks like one is left alone.
Does it work on scanned PDFs?
It can only find text the PDF actually contains. In a scan the words are pictures: the tool names the pages that have no text, and if no page has any it stops and says so. Put the file through OCR PDF first, then redact the result.
Will it find everything?
No, which is why it shows you the list first. It knows data by its look, and a name, an address, a date of birth or an account number that is not an IBAN has none: type those as words. A phone number of eight digits written without a space, and anything broken across two lines or drawn as a picture, is missed. Read the result before you share it.