Guide
How to Redact a PDF Online (Permanently Remove Sensitive Data)
Covering names, account numbers, and personal data in PDFs so the content is truly gone — not just hidden under a rectangle.
By Sorawi Tools Team · Published July 1, 2026
Redaction Is Not the Same as Hiding Content
Dragging a white rectangle or a highlighter over text in a PDF viewer is not redaction. The text is still inside the file; the rectangle is only drawn on top of it. Anyone who removes the rectangle, selects the covered area, or runs a PDF text extractor can recover the original characters as easily as if you had done nothing. This confusion is dangerous precisely because the failure is silent. The document looks perfectly clean, and nobody discovers the leak until the damage is already done. Courts, regulators, and security teams have all been caught out by documents that appeared redacted but were merely masked, sometimes years later when an ordinary search on the file surfaced everything that was supposed to be hidden. Real redaction is permanent destruction: it replaces the underlying content so that no tool, however determined, can pull it back. That is why professional redaction tools do not paint over the text. They burn the pixels out. The only safe redaction removes the sensitive information at the content level, not just the visual layer. That distinction is the whole point of this guide, and it is why the PDF Redactor bakes black boxes into the image of the page rather than adding annotations on top of it.
When You Would Redact a PDF
Redaction comes up whenever a document contains personal or confidential information that must not travel with the rest of the file. A contract or invoice you are forwarding to a third party may carry the other side's bank details, rates, or phone numbers that the new recipient should never see. Medical records and referral letters contain names, dates of birth, and diagnoses that must stay private even when a doctor wants to share part of the file with another practice. Court filings frequently reference children, victims, settlements, or social security numbers that have to be obscured before a document goes to the other party. Compliance rules force the same behavior in business: GDPR gives people the right to have their data removed from copies of documents you hold, and many industries have internal rules about sharing customer information between departments. Even routine paperwork benefits from redaction — resumes carrying an old address, internal memos with project budgets, equipment lists that name contractors, and screenshots of dashboards before a webinar are all cases where a small slice of the file is the only problem. The common thread is that you need to share a document, but a specific part of it must be illegible to everyone who receives it.
How to Redact a PDF with the PDF Redactor
The PDF Redactor renders each page as an image, lets you paint solid black boxes over everything sensitive, and exports the cleaned result. Pages are rendered at a crisp zoom, so text stays legible after export while the redacted areas become flat black. You work page by page, which forces you to look at every page rather than assuming a spot check is enough.
- 1Open the PDF Redactor tool and drag your PDF onto the drop zone
- 2Wait for the first page to render as a drawing surface
- 3Drag a box over every sensitive area, using Undo to fix mistakes
- 4Click Export Redacted Page and download the PDF or PNG output
- 5Step to the next page and repeat until every page is checked
- 6Delete the unredacted original once you have verified the export
What Happens to the Hidden Text
The PDF Redactor renders each page, then overwrites the pixels under every box with solid black before you export. When you download the PDF version, the tool wraps the redacted image into a genuine PDF whose page size matches the original. Because the pixels under each box are destroyed, the text no longer exists in the exported file in any recoverable form — it is gone, not merely covered. This pixel-level approach is what separates real redaction from annotation: no software can select or extract characters that were overwritten at the pixel level, because there is nothing left to extract. The trade-off is that the exported page is an image, so its text is no longer selectable or searchable. For most redaction workflows that is a feature, not a bug, because the document is meant to be read by a person rather than machine-parsed. If you need the text layer preserved everywhere except the redacted parts, keep the original for internal use and share only the redacted export. Either way, the file you hand out contains no trace of what was under the boxes. This is a meaningful step up from tools that simply draw a rectangle on the original text layer, which leave the characters behind waiting to be selected.
Common Mistakes That Compromise Redaction
Most redaction failures come from treating the task as visual rather than destructive. Painting a white box over text in a viewer leaves the text selectable; the fix is to redact at the pixel level the way this tool does. Checking only the first page is another classic error — sensitive information often sits on the last page in a signature block, an appendix, or an attached schedule. Forgetting that many PDFs carry an invisible OCR text layer can expose content that looks blank on screen, so a redacted scan is only as safe as the pixels that actually got covered. Document properties and metadata can leak information too, so strip or inspect them when the file will change hands. Boxes that are too small leave clipped letters peeking out at the edges, which turns a redaction into a hint for anyone determined to read it. Boxes that are too large waste the document and can redact content you actually wanted to share. Another frequent slip is redacting a re-exported version instead of the original, so that the copy still contains the underlying text. The rules that cover all of these: treat redaction as destruction, check every page before exporting, size boxes generously, and never trust the preview alone. A redacted document is only as strong as its weakest page.
How to Verify Your Redaction
Before you share a redacted document, verify it the way an attacker would. Open the exported PDF and use Select All followed by Copy, then paste into a text editor — any recovered text appears immediately. Run the search function of your PDF reader for a string that should be gone, such as an account number or a last name, and confirm there are zero matches. You can also drag the file into a plain-text extractor; if the sensitive string shows up anywhere, the redaction failed and you need to redo that page. Check the boxes visually as well: each one should be opaque black with no characters bleeding out of the edges, and nothing faint surviving inside. The PDF Redactor renders pages at a crisp zoom and exports at the rendered resolution, so edges stay clean as long as your boxes are a little larger than the text they cover. Test at least one file end to end before you rely on the workflow, so you know what your output looks like in practice. Add this verification pass to your routine and you will never ship a document that looks clean but still contains its secrets. It takes a minute, and it is the difference between redacting and hoping.
Privacy and Handling the Original
The PDF Redactor runs entirely in your browser, so your document is never uploaded anywhere. The PDF is parsed locally, the boxes are drawn in memory on your device, and the only files that leave your computer are the ones you download yourself. That matters for medical records, settlement agreements, and client documents, where even routing a file through a third-party server can create a compliance problem. One decision stays with you: what to do with the original. If the document contains information you are legally required to keep, such as an original signed contract, store it somewhere access-controlled and share only the redacted export. If the file is a copy you no longer need, delete it and empty the trash, because the original is still fully readable and is now the weak link. A common error is keeping the unredacted version in the same folder as the redacted one, where a mistaken upload or a wrong attachment can send the wrong file. Naming the redacted export clearly, such as adding a redacted suffix, reduces that risk. Treat the original like the sensitive document it is, and the redaction effort is not wasted.
Redaction vs. Encryption vs. Deletion
Redaction is one of three ways to protect sensitive content, and they solve different problems. Encryption scrambles a file so that anyone without the key sees gibberish — but the information is still fully present, and the file is only protected while it is encrypted. Password-protecting a PDF is encryption, not redaction: the recipient who knows the password sees everything, and the document stops being protected the moment it is decrypted. Deletion removes the whole file, which works only when you do not need to share any of it. Redaction sits in between: it lets you share the document while permanently removing only the parts that must not be seen. That makes it the right choice whenever a file has to change hands but a specific slice of it cannot. The differences matter in practice. A contract that is password-protected is still one bad share away from leaking its bank details; a contract that has been redacted has none to leak. Many organizations apply both layers — redact the sensitive content, then encrypt the redacted file for transport — because each layer closes a different gap. Understanding which protection you are applying stops you from assuming a redacted-looking file is safe when it is merely hidden.
Working with Scanned and Image-Based PDFs
Not every PDF starts as text. Scanned documents are images of pages, sometimes with an invisible OCR text layer added by the scanner software and sometimes without one. Both cases change how you should redact. If the scan has no text layer, the only content is the picture itself, so painting over the pixels removes it completely — the PDF Redactor handles this case naturally, because it works on the rendered page image. If the scan does have an OCR layer, the picture is only half the story: the hidden text under a painted box can still be extracted by a determined tool, which is why redacting the pixels rather than covering the layer is the safe move. A handwritten document with no OCR layer is the simplest case — the redaction box removes what is there and nothing is hidden underneath. For documents that mix printed and handwritten content, such as a signed application form, check both kinds of content on every page. The same verification applies as to any other PDF: after exporting, search for strings that should be gone and confirm nothing survives. Image-based documents make this check particularly important, because it is easy to assume a scan has no hidden text when the scanner quietly added an OCR layer.
PDF Redactor
Draw black boxes over sensitive content in PDF pages and export the redacted page as an image.
