Redact a PDF So the Text Is Actually Gone
Most PDF 'redactions' that make the news failed the same way: a black rectangle drawn over text that was still there underneath, one copy-paste away from exposure. This tool redacts properly — marked regions are removed and pages are flattened to pixels, then an automatic validation confirms the redacted text can no longer be extracted from the output. Everything runs in your browser; the document is never uploaded.
The actual tool runs in our ad-free secure workspace — nothing on this page processes your file.
Open Redact a Document →What this tool does
- Lets you mark redaction areas by drawing boxes or by searching the PDF's text for names, numbers, or phrases.
- Flattens redacted pages to pixel images, so there is no hidden text layer left to copy out from under the black boxes.
- Automatically validates the exported file, verifying that the strings you redacted cannot be extracted from it.
- Strips document metadata (author, dates, software) in the same pass, closing the side channel most people forget.
- Shows a full list of redaction areas before export so you can review every mark.
Your privacy on this tool
Stays on your device
- The PDF is rendered, marked, redacted, and re-exported entirely in your browser.
- Text search for sensitive terms runs on-device — your search terms and matches are never transmitted.
- The validation step also runs locally, on the exported file, before you share it anywhere.
Reaches our server: nothing
This tool makes no upload. Your content is processed entirely in your browser.
How to use it
- Open the redaction tool at /app/redact-document.
- Load your PDF and page through it, drawing boxes over anything sensitive.
- Use text search to find every occurrence of a name, account number, or phrase — matches can be redacted in bulk.
- Review the list of redaction marks, then export.
- Check the validation result confirming the redacted strings are not extractable from the output.
- Share the redacted copy; keep the original somewhere safe if you need it for your records.
Common uses
- Producing documents for legal discovery or a records request with privileged material removed.
- Redacting account numbers from bank statements needed as proof of income or address.
- Removing patient identifiers from medical records used for a second opinion or an appeal.
- Blacking out other parties' details in a contract before showing it to a new counterparty.
- Publishing a report with sources' names removed — verifiably, not cosmetically.
Supported formats
- PDF (input and output)
- Also accepts PNG, JPEG, and WebP images
Works in all modern browsers with the Canvas API; long documents redact most comfortably on desktop screens.
Limitations & security notes
Limitations
- Flattening converts redacted pages to images: text on those pages can no longer be selected or searched by recipients, and file size may grow. That trade-off is what makes the redaction real.
- Scanned PDFs have no text layer, so text search cannot find matches in them — redact scans visually, box by box.
- Redaction removes what you mark; it cannot know what you missed. Careful review is still a human job.
- Context can defeat redaction — 'the only cardiologist in [small town]' identifies someone with the name blacked out.
- SecretPNG is in beta and has not been independently audited.
Security notes
- Drawing a black rectangle in an ordinary PDF editor hides nothing — the text remains in the file, recoverable by select-and-copy. Flattening to pixels is the difference between concealment and removal.
- The automatic validation is your receipt: it attempts to extract the redacted strings from the export and passes only when they are gone.
- Metadata is stripped alongside content because a redacted document that still names its author in the properties defeats the purpose.
- Use text search for anything that repeats — account numbers and names have a way of appearing on more pages than you remember.
- For maximum caution with life-critical redactions, also print the exported file to PDF and inspect it fresh.
Frequently asked questions
- Why is drawing a black box in Acrobat or Preview not enough?
- Because in most editors that box is a shape placed on top of the text — the text object still exists underneath and comes out with select-all or a text extractor. Real redaction removes the content and, in this tool's approach, flattens the page to pixels so there is no underlying layer at all. Court filings and government releases have repeatedly been burned by the difference.
- What does the validation step actually check?
- After export, the tool runs text extraction against the redacted file and searches for the exact strings you redacted. If any redacted string can still be pulled from the output, validation fails and tells you — instead of letting you ship a document that only looks safe.
- Can recipients still search and copy text in the redacted PDF?
- Not on pages that contain redactions — those pages are flattened to images, which is precisely why the hidden-text problem cannot occur. Pages without redactions are unaffected. If recipients need selectable text on redacted pages, that requirement is fundamentally at odds with pixel-level redaction.
- Does this work on scanned documents?
- Yes, with one caveat: scans have no text layer, so the search feature has nothing to search. You redact scanned pages visually by drawing boxes. The flattening and metadata stripping work exactly the same, and validation still confirms no extractable text remains.
- Is my document uploaded for processing?
- No. Rendering, redaction, flattening, validation, and export all happen inside your browser. For documents sensitive enough to redact, that is the only architecture that makes sense — the material never transits anyone else's machine, including ours.
Related tools
Last reviewed: 2026-07-14Open Redact a Document
SecretPNG is in beta and has not been independently audited. Security status.