QuickMergeRedact a PDF — Black Out Text Permanently
Short answer: To permanently redact sensitive PDF text, select the information to remove and export a new document where the underlying text is removed, not merely covered. QuickMerge keeps the document in your browser while it works.
Automatically finds SSNs, credit-card numbers, emails, phone numbers, and your own keywords — then blacks them out permanently. The redacted copy has no text layer to un-hide. All in your browser.
How do I permanently redact text in a PDF?
Drag & drop here
or click to browse
Findings
True Redaction — Not a Black Box You Can Copy-Paste Under
The redaction horror story is always the same: someone draws a black rectangle over text, the text remains underneath, and a journalist or opposing counsel simply selects and copies it. This tool does redaction properly. First it detects sensitive data by pattern — Social Security numbers, credit-card numbers (validated with the Luhn checksum so random digits don't false-positive), email addresses, phone numbers, IBANs, plus any names or keywords you add. You review every finding. Then it redacts by rasterization: each page is re-rendered as an image with black boxes burned in, and a new PDF is built from those images. The original text layer is gone entirely — there is nothing to select, copy, or un-hide.
Detection and redaction both run in your browser, which matters precisely because the documents you redact are the ones you least want uploaded. Everything runs on your device via open-source browser libraries — your files are never uploaded. You can verify it yourself.
Next steps
Before this step
- make a scanned PDF searchable first — text has to be findable before it can be found and removed.
Related tools
- remove a signature from a PDF — a different job: deleting an object rather than removing text.
- verify nothing was uploaded — especially important for documents you are redacting.
What to do with the result
- strip the metadata too — names and software details survive redaction of the page content.
- encrypt the redacted file — before it leaves your device.
Browse all PDF tools, or see every QuickMerge tool.
Wondering how big a file this can take? We measured it: see Redact PDF on the measured limits page, or the browser support matrix.
Frequently Asked Questions
Yes. Unlike drawing rectangles over text, this tool rebuilds the PDF from page images with the black boxes burned into the pixels. The output has no text layer at all — copy-paste, search, and text-extraction tools recover nothing.
The redacted copy behaves like a high-quality scan: it is no longer searchable or selectable (that is the point), and the file size may change. Keep your original for reference and share only the redacted copy.
SSNs (123-45-6789), credit/debit card numbers (Luhn-validated), email addresses, phone numbers, and IBANs — plus any custom words or names you type in, such as a client name or project codename.
Scans have no text layer to search. Run the file through our OCR tool first, or use the custom-keyword field after OCR. You can also verify any PDF's hidden text with copy-paste before trusting a redaction from any tool.
Mark the region here and download the result. The text inside it is removed from the file rather than covered, so it cannot be recovered by selecting or extracting.
Not from the downloaded file — the characters are gone. Keep in mind that your original still contains everything, so control that copy too.
It can detect common patterns — Social Security numbers, credit-card numbers, email addresses and phone numbers — and you can add your own keywords. Always review the matches before downloading; pattern detection is an aid, not a substitute for reading the document.
A scan is an image, so there is no text layer to remove and the marked area is obscured visually. For scans, verify by exporting the page and checking that the covered region is genuinely unreadable. Where the file has been through OCR, a text layer exists and can be redacted properly.
Yes. The document is processed in your browser and never uploaded, which is the point — sending a file full of Social Security numbers to a server to have them removed defeats the exercise.
Mark the regions you want gone — the tool can find common patterns for you, and you can add keywords — then download. The characters inside each region are taken out of the file rather than covered, so selecting or copying that area afterwards returns nothing.
This page does it at no cost and with no account, because the work happens in your browser rather than on a metered server. That is also why it is the right choice for the job: sending a document full of Social Security numbers to a server to have them removed hands them over first.
Open the downloaded file and search for something you redacted with Ctrl+F or Cmd+F. No result means the text is genuinely removed. A highlighted hit under a black box means the tool that produced it only drew a shape, and the data is still in the file.
You cannot, if it was done properly — the text no longer exists in the file. If a black box can be moved or deleted to reveal text underneath, that was never a redaction; it was a drawn shape, and the document should be treated as if it had never been redacted at all.
A black rectangle is safe only when the underlying data is removed
Visual covering can leave searchable text beneath the mark. Always inspect the exported PDF rather than assuming the preview is the final security boundary.
Search beyond obvious identifiers
Review names, addresses, account numbers, signatures, barcodes, document properties and repeated headers. Automated patterns help find common email, phone and card formats, but context-specific identifiers still need human review.
Test the finished file
Try selecting and copying text around every redaction, search for removed terms, and inspect the page at high zoom. Open the download in a second viewer. For scans, check that the redaction covers the image pixels as well as any hidden OCR text layer.
Use approved workflows for high risk
Court filings, access-to-information releases and regulated records may require formal redaction software, audit logs and independent review. Preserve an unredacted master under appropriate access controls and distribute only the verified derivative.
Redaction checklist reviewed 11 July 2026 · Corrections and methodology
Why drawing a black box over text is not redaction
This is the distinction that causes real data leaks, and it is worth being precise about what happens here.
A black rectangle leaves the text underneath
Drawing a filled shape in a PDF editor hides text visually while leaving it in the file. Anyone can select, copy or extract it, and this is exactly how published documents have leaked names and figures that were believed to be hidden.
This tool removes the text layer
The characters within a redacted region are taken out of the document, not painted over. After download, searching or copying from that area returns nothing, because there is nothing left to return.
Check the metadata as well
Redacting the page body does not touch document properties, and author, software and revision fields sit outside the page content entirely. Run the metadata tool over anything sensitive before it leaves your hands.
How to redact in a PDF without leaving the text behind
Redaction has a specific failure mode that looks like success on screen. Knowing what to check is most of the skill.
Work through the document, not just the obvious page
The pattern detection here finds Social Security numbers, credit-card numbers, email addresses and phone numbers, and takes your own keywords as well. Use it as a sweep, then read the document. Detection catches formatted identifiers reliably and misses the things that identify a person in prose — a job title in a small team, an unusual address, a date that only fits one person. Those need eyes.
Verify by searching the downloaded file, not the preview
This is the check that would have caught most published redaction failures. Open the file you downloaded, press Ctrl+F or Cmd+F, and search for something you redacted. If the search finds nothing, the text is genuinely gone. If it highlights a hit under a black rectangle, whatever produced that file drew a box rather than removing the characters.
A scan has no text layer, so redaction means something different
On a scanned page there is nothing to delete — the page is an image, and covering an area is all that can be done. That is adequate only if the covering is flattened into the image itself. If the document went through OCR at some point, a real text layer exists and can be removed properly. Check which kind of file you have before assuming a black box is enough.
The page body is not the whole document
Redacting the visible text leaves document properties untouched — author, the software that produced it, revision history, and in some workflows the original filename, which on a legal or medical document is often the leak. Run the metadata tool over anything sensitive before it goes out, and note that your own original still contains everything.