Why “drawing a black box” is not redaction
The most common redaction mistake is also the easiest to make. Someone opens a PDF, draws a black rectangle over a name and saves it. On screen the name is gone. In the file it is still there, under the rectangle. Anyone can select the text and copy it, search for it, or simply delete the rectangle. Court filings, government reports and company documents have all leaked private details this way.
Real redaction removes the information from the file. This tool does that by rebuilding every page that has a black box as a flat image with the boxes burned in. The page’s original text, fonts and drawing commands are deleted. What is left is a picture of the page in which the covered areas are solid black pixels. Nothing remains underneath.
How to redact a PDF
- Choose your PDF. Its pages appear below the search panel.
- Find text automatically (optional). Type names, account numbers or phrases, one per line, and tick any presets: email addresses, phone numbers, long numbers such as ID, card and account numbers, or dates. Click Find and mark. Every match on every page gets a black box.
- Draw boxes by hand. Drag across anything else that must go: a signature, a photo, a handwritten note, a stamp. Remove a wrong box with its red ×.
- Choose the image quality for redacted pages. 200 dpi keeps normal print sharp; use 300 dpi for small print.
- Click Redact and save and download the file, which ends in “-redacted”.
- Check it. Open the file, try to select or search a redacted word, and look at every page.
What exactly gets removed
| Content | Removed? |
|---|---|
| Text, images and drawings under a black box | Yes, the whole page is rebuilt as an image |
| Other text on a redacted page | Becomes part of the image (not selectable) |
| Pages without any box | Unchanged, text stays selectable |
| Comments, form fields and links on redacted pages | Yes |
| Document title, author, keywords, XMP metadata | Yes, with “remove hidden data” ticked |
| File attachments and scripts | Yes, with “remove hidden data” ticked |
| Accessibility tags (they can hold copies of the text) | Yes, with “remove hidden data” ticked |
| Bookmarks | Optional, as their titles may contain names |
Only pages you redact lose their text layer. A 40-page contract with two redacted names on page 7 keeps 39 fully searchable pages. If you need page 7 searchable again, run OCR PDF on the result. OCR reads only what is visible, so the black areas stay empty.
A worked example
An HR manager must share a 12-page disciplinary report with an outside adviser, without the employee’s name, staff number or personal email.
- She types the name (“Imran Qureshi”) and the staff number (“EMP-20417”) and ticks Email addresses.
- Find and mark places 23 boxes on 9 pages.
- She scrolls through, removes one box that caught the company’s general HR email (which may stay), and draws one more box over a handwritten signature on page 12.
- At 200 dpi, the 1.1 MB report becomes 3.4 MB: nine pages are now images. She runs Compress PDF to bring it back under 2 MB, then Protect PDF to add a password before emailing it.
Search tips
- Search for variations. “Imran Qureshi”, “I. Qureshi”, “Mr Qureshi” and “Qureshi” may all appear. Adding just “Qureshi” catches most of them.
- Searches ignore case and treat any run of spaces or line breaks as one space, so a name split across two lines is still found.
- Presets cast a wide net. “Long numbers” also marks invoice numbers and dates written without separators. Review before saving.
- Scans have no text. On a scanned page the search finds nothing, so draw boxes by hand.
Redacting here vs typical online redaction tools
| This tool | Typical online redaction tools | |
|---|---|---|
| Your unredacted file is uploaded | Never | Yes, to the provider’s server |
| Text truly removed | Yes, pages are rebuilt | Varies; some only draw boxes |
| Search-based redaction | Free | Often a paid feature |
| Removes metadata and attachments | Yes | Varies |
| Account or limits | None | Common |
Uploading a document so that someone else’s server can remove the sensitive parts defeats much of the purpose. Here, the unredacted file never leaves your device.
Limits
- Redacted pages become images. Text on them is no longer selectable and the file usually grows. That is the price of a guarantee that nothing is left underneath.
- Hiding is not the same as redacting. Edit PDF whiteout and Crop PDF both hide content but leave it in the file. Use this tool whenever the information must be gone.
- You are responsible for the final check. Automatic search can miss spellings you did not think of. Read every page of the result before sharing it.