How to redact a PDF safely instead of drawing a black box
Remove sensitive content from the document itself, verify the result and avoid common visual-redaction mistakes.
Covering text is not always redaction
A rectangle placed over text may hide it visually while the underlying characters remain selectable, searchable or recoverable. Changing text color or placing an image above it has the same weakness.
A proper redaction workflow removes the selected underlying content and produces a new sharing copy. Keep the untouched original in a protected location.
Identify every occurrence
Search for names, account numbers and repeated identifiers, then inspect headers, footers, comments, attachments and scanned images manually. Search alone cannot find text that exists only as pixels.
Consider indirect identifiers. A date, job title or case number can reveal a person even when their name is removed.
Apply and export
Mark only the areas that must be removed and read the tool's limitations. Export a new file rather than overwriting the source.
Scanned pages need image-aware redaction. Complex forms, layers or embedded files may require specialist desktop software and an organizational review process.
Verify like a recipient
Open the redacted copy in a different PDF viewer. Try selecting text around the redacted area, search for the removed terms and copy nearby content into a plain-text editor.
Zoom in to check edges, and review document properties, comments and attachments where relevant. Verification is part of redaction, not an optional final polish.
Share the minimum necessary file
Extract only the pages the recipient needs before redaction when practical. Use a clear filename and send the redacted copy, never the working source.
For legal, regulatory, medical or high-risk disclosure, follow the approved procedure of the responsible organization rather than relying only on a general-purpose browser tool.
How MyFileNest Redact PDF works
Redact PDF removes a broad band of each page you list. You enter the pages, for example 1, 3-5, choose whether the band sits at the top, centre or bottom of the page, and set its height between 5% and 60% of the page. The band spans almost the full page width, which suits address blocks, letterheads, signature areas, account summaries and similar sections that run across a page.
The tool then renders every page of the document as an image, paints the band black on the pages you selected, and builds a brand-new PDF from those images. Nothing that sat under the band survives, and because the new file is built from scratch, the original document properties are not copied into it. The trade-off is that all pages in the redacted copy become images, so their text is no longer selectable or searchable. Run OCR PDF on the redacted copy if you need searchable text again.
Because the band is a fixed region rather than a word-by-word selection, it is best suited to information that sits in a predictable area. For a name scattered through several paragraphs, a dedicated desktop redaction tool that removes individual words is the better choice.
Famous kinds of redaction failure
Redaction mistakes in published documents have repeatedly made the news. They tend to fall into a few patterns, all of which the verification steps below are designed to catch.
| Failure | What went wrong | How to avoid it |
|---|---|---|
| Black box over live text | Text still selectable and copyable underneath | Use a tool that removes content, then test by selecting and copying |
| Highlighting in black | A highlight annotation was mistaken for redaction | Check the comments or annotations panel; flatten or remove content |
| Missed occurrences | Name redacted on page 2 but not in a footer on page 9 | Search the whole document for each term before and after |
| Metadata and properties | Author or title fields still contain the name | Check document properties; rename and rebuild if needed |
| Indirect identification | Name removed but job title and date identify the person | Review context, not just the obvious terms |
Step-by-step: a safe redaction routine
- Make a working copy and keep the original somewhere secure.
- Extract only the pages the recipient needs with Extract PDF Pages.
- Search the working copy for every name, number and identifier to remove, and list the page of each occurrence.
- Open Redact PDF, enter the pages, choose Top, Center or Bottom and set the band height so it fully covers the sensitive area with a small margin.
- Export the redacted copy.
- Open it in a different viewer. Check that the band covers everything it should on every listed page, and that nothing sensitive sits outside it.
- Check document properties and the file name for personal information.
- Send only the redacted copy.
Worked example: publishing a tenancy dispute decision
A residents' association wants to share a two-page decision letter with members without revealing the tenant's name, flat number and phone number, which all sit in the address block at the top of page 1. They choose page 1, the Top region and a band height of about 25%, enough to cover the letterhead's address block with a margin. Reading the redacted copy, the address block is solid black, the body of the letter remains readable, and the surname appears nowhere else because the letter refers only to "the tenant". They also rename the file from the tenant's surname to decision-letter-september.pdf before sharing it.
What to redact: a practical checklist
Different situations require different redactions, but the same categories come up again and again. Work through this list for every document you prepare for disclosure.
- Direct identifiers: names, addresses, phone numbers, email addresses, dates of birth and signatures.
- Official numbers: national insurance or social security numbers, passport and licence numbers, tax references and patient or case numbers.
- Financial details: account and card numbers, sort codes, salaries and amounts that identify a person.
- Third parties: names and details of people who are not the subject of the request, such as witnesses or other customers.
- Indirect identifiers: job titles, unusual events, small-team references or dates that point to one person.
- Images: faces, handwriting, number plates and house numbers in photographs or scans.
Redaction for data subject and freedom of information requests
Organisations responding to requests for personal data or public information often need to release documents while protecting other people's details. In many legal systems, including data protection laws in the UK and EU, a person asking for their own data is entitled to it, but information about other people may need to be removed. Freedom of information regimes have their own exemptions for personal, commercial and security-sensitive information.
Keep a log of what was redacted and why, especially when the redaction relies on a legal exemption. If the decision is challenged, the log and the unredacted original, kept securely, show how each redaction was justified. For high-volume or legally sensitive work, follow your organisation's approved procedure and tools; this guide covers the technical side, not the legal judgement.
Redacting scanned documents and images
Scanned pages contain no selectable text, so search tools cannot find the information you need to remove. Read each page carefully, or run OCR on a working copy to search it, and then redact the areas on the scan itself. Because MyFileNest Redact PDF flattens pages to images, the black band becomes part of the page picture, which is exactly what is needed for scanned material.
For standalone photos and screenshots, crop out sensitive areas where possible, as cropping removes them completely. Blurring and pixelation can sometimes be partly reversed or guessed for small text such as numbers, so a solid black box is safer for anything genuinely sensitive.
Quick answers
Can I undo a redaction? Not in the redacted copy; that is the point. Keep the original securely in case you need to redo the work.
Why is the redacted PDF no longer searchable? Redact PDF flattens pages to images so nothing survives under the black band. Run OCR PDF on the redacted copy if you need searchable text again.
Is the document uploaded? No. Redaction runs in your browser, which suits confidential material.
Can I redact several areas on one page? The tool removes one band per run; run it again on the redacted copy for a second area, keeping in mind each run re-renders the pages.
Before you start
Set aside uninterrupted time for redaction. Most failures happen when people rush, skip the verification step or work on the wrong copy. Close unrelated files, label your working copy clearly, and complete one document fully, including checks, before starting the next.
Summary
Work on a copy, extract only the pages needed, identify every occurrence of sensitive information including indirect identifiers, and remove it with a tool that deletes the content rather than covering it. Then verify the result as a recipient would, in a different viewer, by searching and checking each redacted area. Keep the original securely, record why each redaction was made when it matters, and share only the redacted copy.
MyFileNest Redact PDF suits information that sits in a predictable band of the page, such as address blocks, letterheads and signature areas, and flattens every page so nothing survives under the band. For word-by-word redaction across complex documents, use dedicated desktop software.