How it works

What it looks for

Emails, phone numbers, card numbers, IBANs, and a set of country IDs are checked for you. Names and addresses are included when they sit on a labeled line, such as Bill to or Address. Money is listed and left unchecked, because a salary or an invoice total is often the reason you are sending the PDF.

India dummy and US dummy are invented files on the tool. Use them before you upload anything of your own. The limit is 25 MB. There is no account.

Do it in this order

1. Drop a PDF on the upload panel, or click a dummy. A scan with almost no text takes longer, because those pages are read with OCR first.
2. Read every row. Uncheck anything the recipient is supposed to see. Leave amounts unchecked unless the figure itself is the secret.
3. Click Hide selected. The server paints a dark rectangle on the checked boxes and rebuilds the PDF from page images. If a hidden value is still extractable as text, you get an error and no file.
4. Zoom the preview, download, and search the download for one value you meant to remove.

Show all, then Hide selected, shows the original file. That download is not a redaction. The field-by-field version is how to redact a PDF properly.

Situations people actually bring here

A recruiter checking a salary needs your name, the employer, the period, and the gross figure, not a national ID or the bank account in the footer. Leave the money unchecked. Sharing a payslip uses the US dummy for that split. A letting agent proving address and income needs a different cut: hide the account number, and still read payee names the patterns miss. See bank statements and personal data.

Hiding, redaction, and the limits

A black rectangle in Preview changes the picture and leaves the letters in the file, so copy-paste still works. Why black boxes do not hide PDF text shows that on the sample SSN. Hide selected is closer to redaction: the download has no text layer, and a dark block sits on the fields you kept. It still misses names in ordinary sentences, handwriting, and odd number formats. Text you did not cover stays readable in the picture, and someone can OCR it later. The whole page is an image, so harmless words are not selectable either, and a digital signature does not survive. If a rule tells you how to redact, follow the rule.

Search the download, try to copy from the block, and open document properties. How to check a PDF is really redacted is that checklist. Removing metadata explains what the new file drops, and the producer name it can still carry.

Scanned pages

Pages that already have a text layer are read from that text, and OCR is skipped. A page under about 40 characters is treated as a scan. The server renders it, writes a temporary PNG, runs Tesseract, and deletes the PNG when that step finishes, including when OCR fails. A misread digit can drop a PAN or an Aadhaar from the list while the number stays visible. An empty list on a page of ID numbers is a failed read. Zoom the preview. The scan guide goes further.

What happens to the upload

The PDF is held in memory for this request. It is not written into a library or a database, and it is not kept after the response is sent. The only file written on purpose is the temporary OCR image, removed when that step ends. Request logs can include the time, path, and status. They do not include the PDF. Cookies and the AdSense disclosure are on the privacy page.