Redact sensitive data in a PDF
Before handing a document to a third party — answering a subject-access request, sharing exhibits in a lawsuit, publishing an attachment — some data has to be gone for good. And a black rectangle drawn over text is not redaction: anyone can copy-paste underneath it. This tool scans your PDF and automatically detects personal data: IBANs, payment card numbers, emails, phone numbers, dates of birth, passport and driving licence numbers, IPv4 addresses, structured national identifiers (French social security numbers, SIREN/SIRET) and person names introduced by a label or honorific ("Name:", "Mr", "Mrs", "born on"…). Tick the items to hide, or target your own keywords — a name in free text, an internal case reference, an address. Redacted zones are then rasterized: the content under the black boxes is truly deleted from the file, not merely covered. Detection and redaction run entirely on your device — exactly what you want for documents sensitive enough to need redaction. Honest limit: the scan covers common French and European patterns, so always finish with a human review and your own keywords for names in running text or case-specific references.
FAQ
Is redaction really permanent?
Yes: unlike a black rectangle drawn on top (reversible with copy-paste), redacted pages are rasterized — the text under the zones no longer exists in the file produced.
What does the automatic scan detect?
IBANs, French social security numbers, payment cards, SIREN/SIRET, emails, phone numbers, birth dates and IPv4 addresses — identified through patterns and checksums, entirely on your device. It also spots passports, driving licences and person names introduced by an explicit label ("Name:", "Beneficiary:"…) or an honorific ("Mr", "Mrs", "born on"…).
Can I redact terms the detection does not know?
Yes: enter your own keywords (a name, an address, an internal reference) — every occurrence is offered for redaction.
Does the detection catch every person's name?
It catches names introduced by a label or an honorific ("Name:", "Mr", "born on"…); a name cited freely in running text must be targeted with your keywords.
Does redaction also remove metadata?
Redaction works on page content; to erase author, dates and file history, follow up with the Remove Metadata tool — the natural complement for a fully anonymized file.
Why redact locally rather than online?
Because sensitive documents are precisely the ones you least want to upload: neither the document nor the detected data ever leaves your browser.
If I make a mistake, can I undo it?
Your source file is never modified: if the redaction goes too far, start over from the original with a different selection.
Is a redacted PDF still searchable?
Redacted pages are rasterized, so the removed text is gone for good — including to search and copy-paste. The rest of the document keeps its normal text layer.
Can I rely on the automatic detection alone?
No — it speeds up the hunt for sensitive data, but legal or public release demands a human review of every proposed zone, plus your own keywords for anything context-specific.
Does detection work on a scanned PDF?
A scan is an image with no text layer, so there is nothing for the detectors to read — run the “OCR” tool first to add the text layer, then relaunch the scan. Keyword redaction works the same way afterwards.
What are the file limits?
One document up to 50 MB and 1,000 pages: detection reads the text layer of every page, so a long file takes a little time — still entirely on your device.
How to proceed
- Drop your PDF — the personal-data scan starts automatically
- Tick the detected items to redact, or type your own keywords
- Review each proposed zone — you confirm before anything is removed
- Download the redacted PDF — selected zones are permanently deleted
- Reopen the file and try copy-pasting under a black box to verify