Why a Black Box Doesn't Redact a PDF (and What Does)
Covering text with a black rectangle changes how a page looks, not what the file contains. Here is how to remove the text for real, and how to check that you did.
A black rectangle drawn over text hides nothing: the text stays in the PDF, and anyone can select, copy or search it. Redacting for real means deleting the content underneath. You can redact the PDF in your browser: the file is processed on your device without being uploaded, and the tool checks the final file before it lets you download it.
Why doesn't a black box hide the text?
A PDF stores text as characters with a position on the page. A black box is a different object: a filled rectangle drawn on top, in another layer. The page looks covered, but the characters are still in the file, exactly where they were. What you see on screen is not what the file holds.
Google's documentation says the same thing. It advises against placing black rectangles over text to hide it, because the text remains in the document, and recommends proper redaction tools plus a check of the metadata.
A real case: Paul Manafort's 2019 court filing
On January 8, 2019, Paul Manafort's lawyers filed a public version of a response to the government's breach submission (United States v. Manafort, the federal court in Washington, D.C., case 1:17-cr-00201, docket entry 471, 10 pages). Some passages were covered with black boxes. Mother Jones reported that day that the redaction had failed, quoting a tweet from David Martosko: "You can copy the text right from underneath the black boxes."
We looked at the public PDF from the docket. Page 6 has six filled rectangles, each one line tall. When we extracted page 6 with pdf.js on October 21, 2026, one sentence came back: "The same is true with regard to the Government's allegation that Mr. Manafort lied about sharing polling data with Mr. Kilimnik related to the 2016 presidential campaign." In the page content, that sentence sits under two of those rectangles: its baselines fall inside them.
Where else the data you thought you covered hides
The text under the box is only the first place to look. A file can keep the same information in other spots, and each one needs its own check:
- the text layer under the box;
- form fields and comments;
- metadata (author, title, XMP);
- bookmarks with the text in their title;
- attachments;
- earlier versions saved at the end of the file (incremental saves);
- embedded page thumbnails;
- the hidden OCR text in a scan.
What else a PDF carries besides the visible text goes through these layers one by one.
How to redact a PDF properly, step by step
- Work on a copy, so the original stays intact.
- Mark every occurrence: search for the value instead of trusting that you spotted it once.
- Use a tool that removes the content, not one that only draws a shape.
- Remove the metadata and attachments, for example with Sanitize PDF or PDF metadata, and flatten any forms with Flatten PDF.
- Check the final file: search with Ctrl+F or Cmd+F, and select and copy over the covered area to see what comes out.
- If the value appears in an image or a scan, look at the page by eye: text search cannot read it.
What Docuboxer's Redact PDF does (and doesn't)
With Redact PDF you draw boxes, or search for text or a regular expression and mark the matches. It also offers ready-made patterns: email, phone numbers (Mexico, Spain and the US), Mexican tax and ID numbers (RFC, CURP), Mexican bank account numbers (CLABE), Mexican social security numbers (NSS), Spanish DNI/NIE ID numbers, US Social Security numbers, IBANs and card numbers checked with Luhn. It removes the text under each box with PDFium and keeps the rest of the document.
It flattens forms and annotations, and rebuilds the document without the document info (author, title), XMP, bookmarks or attachments. Before the download, it checks the final file and blocks it if it finds the marked text or leftover structure. On scanned pages, it removes the pixels under the box, so the file can grow. All of it happens in your browser, with no upload.
The limits are real. It does not read text inside images, because there is no OCR, so scanned pages need your eyes. The size of a box shows roughly how long the hidden value was, and a 2022 study found that the positions of the characters left around a redaction can help guess short names. If a PDF has a password, open it first with Unlock PDF.
Frequently asked questions
Does a black box over text redact a PDF?
No. The box is drawn on top and the text stays in the file, where it can be selected, copied or found with search. To remove it you need a redaction tool that deletes the content underneath.
How do I check that a PDF is properly redacted?
Open the final file, search for the value with Ctrl+F or Cmd+F, and select and copy the covered area into another document. If it shows up, it is still there. Check the metadata too and, for scanned pages, look at the page itself.
Does printing and rescanning a PDF work as redaction?
It removes the text layer, but you get images with no selectable or searchable text, a bigger file, and OCR on the scanner can create text again. A real redaction removes only what you mark and keeps the rest of the text.
Can a redaction leak anything after the text is deleted?
Its shape can. The width of the box hints at how many characters were there, and a 2022 study showed that the positions of the surrounding letters can help guess short names. For very sensitive data, cover the whole phrase.
Does Docuboxer's Redact PDF upload my file?
No. The PDF is processed in your browser with the PDFium engine and is never sent to a server. You can confirm it in the Network tab of the developer tools while you use it.
Can I redact a scanned PDF?
Yes, by drawing a box over the area: the pixels underneath are removed. Text search cannot find data inside an image because the tool does not run OCR, so check every scanned page by eye.
Redact your PDF without uploading it
Remove the text under each box and check the file before you download it. In your browser, free, no signup.
Open Redact PDF →Related tools
- Redact PDF: Removes the text under each box and verifies the result.
- Sanitize PDF: Strips active content, attachments and metadata, with a report.
- PDF metadata: Read, edit or delete author, software and dates.
- Flatten PDF: Lock forms and annotations so nobody can change them.
- Unlock PDF: Open a password-protected file before you redact it.
Sources
- United States v. Manafort, D.D.C. 1:17-cr-00201, entry 471: REDACTED Response to the Government's Breach Submission, filed January 8, 2019 (CourtListener) (accessed October 21, 2026).
- Mother Jones, Inae Oh, Paul Manafort's Lawyers Tried—and Failed—to Redact This Sensitive Legal Filing (January 8, 2019) (accessed October 21, 2026).
- Google Search Central, Keep redacted information out of Google Search (last updated December 31, 2025) (accessed October 21, 2026).
- Bland, Iyer and Levchenko, Story Beyond the Eye: Glyph Positions Break PDF Text Redaction (PoPETs 2023; arXiv 2206.02285, June 5, 2022) (accessed October 21, 2026).