Why a black box is not a redaction
In a PDF, text is stored as characters with a font and a position. A black rectangle drawn over it with an editor, a comment tool or a highlighter in black is a second object on top. The page looks redacted, but the characters are still there: anyone can select them, copy them, find them with search, or move the rectangle away. The NSA's 2005 guide Redacting with Confidence names covering text with black rectangles or black highlighting as the most common mistake, and says this works for paper but not, in general, for documents shared as files.
A real redaction removes the content. Adobe Acrobat's own redaction command is described in its JavaScript reference as removing all underlying content under the marks. This tool reaches the same goal in a simpler, very thorough way: every page that has a box is redrawn as a picture with the boxes painted into its pixels, and nothing else of the old page is kept.
What happens to a redacted page
| Page with a box | Page without a box | |
|---|---|---|
| What is saved | One picture of the page, at 150, 200 or 300 dpi, with the boxes burned in | The page as it was |
| Text | Gone: nothing to select, search or copy | Selectable and searchable |
| Fonts, images, drawings | Removed; only the picture is left | Unchanged |
| Links, comments, form fields | Removed (comments and filled fields still show in the picture as they looked) | Unchanged |
| Page size, bookmarks to the page | Unchanged; bookmarks keep working | Unchanged |
The tool also takes out the page's thumbnail and the document's tags, the structure that screen readers use, because the tags describe text that is no longer there and can carry copies of words as alternate text. When it saves, it leaves out every object the document no longer uses, so the old page content is not hidden somewhere in the file. Then it opens the new file and reads its text the way a PDF reader does. If a redacted page still had any text, the file would not be saved.
If you prefer the whole document to look the same, choose Images too under Pages without boxes. Every page then becomes a picture.
Find emails, phone numbers and words
Find and mark searches the text of every page and puts a box over each match. Email addresses are found by the pattern name@domain.tld. Phone numbers are groups of digits separated by spaces, dots or dashes, optionally with a + and a country code or an area code in brackets, with 7 to 15 digits in total; dates like 2026-10-23 are skipped. Other numbers written the same way, such as some order or book numbers, are marked too; untick them in the list of matches if they may stay. Words and phrases you type are found regardless of upper and lower case, with Whole words only so that "Ann" does not mark "Annual".
A PDF tells where each piece of text starts and how long it is, but not where every single letter sits, so the tool estimates the letters and then looks at the drawn page to stretch each box to the end of the word it covers. Still, look through every page before you save: every box can be moved and resized, unticked matches are left alone, and anything the search did not catch can be covered by hand. Scanned pages hold pictures of text, which the search cannot read; draw the boxes there yourself.
After saving, the check lists every covered word or number that still appears somewhere else in the new file: on a page without a box, in a bookmark title, in a comment or in a form field. Then you can decide whether to cover that too.
The rest of the file
Redaction is about the pages, but a PDF can repeat their content elsewhere. By default the tool also removes the document properties and XMP metadata, where a title or author can name what you black out, and attached files, which it cannot look inside. Bookmarks stay unless you tick them, because they help readers; the check tells you when a bookmark title contains a covered word. To look at the metadata first, use the PDF Metadata Editor; to take comments out of the unredacted pages, use Remove PDF Annotations.
Good to know
- Redacted pages are pictures, so the file usually gets larger and small print can look softer than before. 300 dpi and the Sharper image option keep it crisp.
- Encrypted PDFs have to be unlocked first with Unlock PDF if you are allowed to.
- A very large page is drawn with fewer pixels when the chosen resolution would go beyond what the browser can draw; iPhones and iPads have a lower limit than computers. The tool tells you when that happens.
- Keep the original in a safe place and only share the redacted copy, which gets "-redacted" in its name.
How to redact a PDF
- 1Drop your PDF on the box or choose it.
- 2Tick Email addresses or Phone numbers, or type names and words, and press Find and mark. Or draw boxes on the page by pressing and dragging.
- 3Go through every page, adjust or add boxes, and untick matches you want to keep.
- 4Press Redact and download. Read the check under the button: it tells you which pages were redacted and whether a covered word still appears somewhere else.
Sources
- NSA Information Assurance Directorate, Redacting with Confidence: How to Safely Publish Sanitized Reports Converted From Word to PDF, report I333-015R-2005 (December 13, 2005): covering text with black rectangles or black highlighting is not effective for documents shared as files, and sensitive content must be removed, not hidden
- Adobe Acrobat JavaScript API reference, Doc methods: applyRedactions, which removes all underlying content under redaction marks
- Arlington PDF Model, PDF Association: the parts of a PDF this tool clears on redacted pages and in the file (page thumbnails, structure tree with alternate and replacement text, metadata streams, embedded files, outlines), as defined in ISO 32000
Frequently asked questions
Is the text under the black boxes really gone?
Yes, on every page with a box. That page is saved as one picture with the boxes painted into its pixels, and its text, fonts, links and form fields are removed. Before the download, the tool opens the new file again and reads its text; if any text were left on a redacted page, it would not save the file.
Why can I no longer select the text on a redacted page?
Because the whole page became a picture, not only the parts under the boxes. That is what makes the hidden text unrecoverable. Pages without a box keep their text, so you can still select and search them, unless you choose to turn all pages into images.
Can I just draw a black rectangle in my PDF editor instead?
Not safely. A rectangle drawn over text in a normal editor or comment tool only covers it: the text is still in the file, and anyone can copy it, search for it or move the rectangle away. The NSA warned about exactly this mistake in its 2005 guide on redacting documents.
Does the search find every email address and phone number?
It finds them in the text the PDF contains, using common patterns: name@domain.tld for emails, and 7 to 15 digits in groups for phone numbers. It cannot read scanned pages, which hold pictures of text, and unusual formats may be missed. Always look through every page and draw boxes for anything the search did not mark.
Is my PDF uploaded?
No. Your browser opens the PDF, searches it, draws the pages and saves the redacted copy on your device. Nothing is sent to a server, which is the point when the document holds things nobody else should see.
Why is the redacted PDF larger?
A page of text takes little space as text and fonts. As a picture at 200 dpi it has millions of pixels. Choose 150 dpi or JPEG for a smaller file, or 300 dpi and the Sharper image option for crisp small print.