Skip to content
Free tool | 100% Local

Redact a PDF — properly

Black out names, account numbers, anything — free, in your browser. The file never leaves your computer, and the export is rebuilt from scratch: no hidden text layer, no leftover metadata.

100% local: the PDF is processed on your device and never uploaded. Auto-detect downloads the text-recognition model (~7 MB) from this site the first time you use it.

How it works

  1. 01

    Drop or pick a PDF — a bank statement, a contract, a scan. Pages render right in the page.

  2. 02

    Drag black boxes over the sensitive parts, or let auto-detect find emails, IBANs, card numbers and API keys on every page.

  3. 03

    Download the result: a flattened, image-only PDF rebuilt from scratch — nothing to copy out from under the boxes.

Why most PDF redaction fails

A black rectangle drawn over text in Preview or a PDF editor usually does not remove the text — the characters stay in the file's text layer, and a select-all, a copy-paste or one pdftotext command brings them back. Court filings and government releases have leaked exactly this way, repeatedly.

This tool cannot make that mistake by construction: the export contains only page images. Every page is rasterized, boxed and re-encoded, so there is no text object left to extract — and blurring instead of blacking out has its own risks, which we measured across 480 cases.

Read the blur recovery study

No upload, no metadata

The documents people redact — statements for a landlord, contracts, medical records — are exactly the files that should never be uploaded to a converter site. This page does everything in your browser: the PDF stays on your device, works offline once loaded, and there is nothing on our servers to leak. And because the output file is rebuilt from scratch, the original's metadata — author, title, editing history — is gone without a separate 'remove metadata' step.

Just need to hide something in a screenshot or photo? Use the image blur tool — same approach, same privacy.

Redacting your screen, not a file?

If you share your screen or record demos, blur sensitive data on the live page before it is captured — one click in the browser, auto-detection included.

Download forGoogle ChromeDownload forMicrosoft EdgeDownload forBrave, Comet, etc.Download forMozilla Firefox

FAQ

No. Rendering, redaction and export all happen in your browser. The file never leaves your device, and the download is generated locally. Auto-detect fetches the text-recognition model from this site once, then also runs entirely on your device.

Because the text usually stays in the file underneath the box. PDF text lives in a separate layer, and covering it changes what you see, not what the file contains — select-all and copy often reveals everything. This tool exports an image-only PDF instead, so there is no text layer left to extract.

Yes, by construction. The exported file is assembled from scratch and contains only the page images — none of the original's metadata, author fields, annotations or editing history is carried over.

Drop the statement in, click auto-detect — it finds account numbers, IBANs, card numbers and addresses-adjacent data like emails and phone numbers — then add boxes for anything personal it missed, and download. The result is safe to send to a landlord or an agency: the numbers are not hidden under the boxes, they are gone.

Yes. Scanned pages are already images, and auto-detect reads them with on-device OCR, so it can find and box a card number even in a photographed document. Very low-quality scans may need a few manual boxes.

No — and that is deliberate. The export is an image-only PDF with no text layer, which is what guarantees nothing can be copied out of it. If you need a searchable file, run OCR over the redacted copy afterwards; the redacted content will not come back.