Black out names, numbers or terms in a PDF - via search terms or drawn directly on the page. The content is truly removed. No upload.
Running locally on your device ...
Rendering page...
Running locally on your device ...
0%
Your files never left your device
Continue with this file
No re-upload neededThese tools accept .PDF - your file is handed over directly.
No. Everything runs in your browser - your file never leaves your device. How this is verifiable
Classic redaction often fails: a black rectangle is placed over the text - but underneath it survives and can be selected, copied or searched. True redaction must remove the content irreversibly. That is exactly what this tool does - in three ways, alone or combined.
Enter terms (e.g. names, case numbers or amounts) - every occurrence is found automatically. Or drag a rectangle directly on the page preview over a spot no text match would catch (e.g. a signature or a photo); multiple rectangles are possible, each removable via its X. Every page with a match or a drawn rectangle is rendered to an image onto which the black bars are burned - the original text of that page is then physically gone. Pages without a match or rectangle stay unchanged.
Five patterns can additionally be auto-detected: email addresses, phone numbers, credit-card-like numbers, US-Social-Security-like formats (123-45-6789) and dates - toggle them independently of the terms list, or combine both. These are pure format patterns, not checksum validation: a credit card number is recognised by prefix and length (not the Luhn check digit), so even an invalid number with a matching format can be caught. Social Security numbers are only recognised in the US format - other countries use different formats. Dates are inherently ambiguous (day/month order varies by country) and can occasionally match unrelated digit sequences.
Everything happens locally in the browser, nothing is uploaded - the most sensitive documents never leave your device. The pattern detection also runs purely client-side as a text match, with no network request at all. The honest price of real security: a redacted page becomes an image, so its remaining text is no longer selectable or searchable.
The practical difference shows on long documents: if you need to remove a name from an 80-page contract, you do not want to hunt down and box every occurrence by hand. The term search here redacts every occurrence at once - combined with the pattern detectors it catches all e-mail addresses or phone numbers along the way. Some well-known free redaction tools offer exactly none of that and make you drag box after box by hand. And because everything runs locally, the most sensitive document in the chain never sits on someone else's server - free, with no account and no page limit.
| Input formats | |
|---|---|
| Auto-conversion | PNG, JPG, WebP, GIF, BMP, TIFF, ICO, SVG, HEIC, ZIP, DOCX, 7Z, DDS, HDR, PCX, WBMP |
| Output format | |
| Batch processing | No |
| Processing | Locally in your browser (JavaScript) |
| File upload | None |
Limitations: Redacts text-based matches (no OCR for scanned image PDFs), detected patterns and/or manually drawn rectangles. Pattern detection checks format only (no checksum): the credit-card pattern matches prefix+length without a Luhn check, the SSN pattern only covers the US format 123-45-6789, and date patterns are ambiguous (day/month) and can occasionally produce false positives. A redacted page becomes an image - the remaining text on that page is then no longer searchable (the price of real security). Review the result before sharing. Open password-protected PDFs first with "Unlock PDF".
No. The affected page is rendered to an image and the bar is burned into the pixels - there is no text layer left from which the content could be recovered.
Yes, draw a rectangle right on the page preview for that - independent of the term search.
No, redaction happens entirely locally in the browser.
Because real security only works if the whole page becomes an image. That is the deliberate price - pages without a match or rectangle stay searchable.
Only if the PDF contains real text (no OCR). For pure image scans, drawing rectangles helps.
It checks format only, no checksum - a credit card number is recognised by prefix and length, not the Luhn check digit; the SSN pattern only covers the US format; and date patterns are ambiguous. Occasional false positives are possible, the tool deliberately favours over-redaction over under-redaction.
Yes. Enter the term - every occurrence across the whole document is found and redacted, whether that is 3 hits or 300. Multiple terms and the pattern detection combine in the same pass.
Yes, completely free - no account, no watermark, no cap on pages or documents. That is possible because the processing runs on your device instead of a server.
Protect PDF with a password · Unlock PDF · Delete PDF pages · Crop PDF · Edit PDF