Redact a PDF
Draw over it, select it, or find every copy of it across the document. Whatever you mark is removed from the file, not covered up. Everything happens on your device.
Drop a PDF here
or browse your device
How to redact a PDF
- Open your PDF. Drop the file onto the page or browse for it. It opens in the editor without leaving your device.
- Mark what has to go. Drag a box over an area, switch to Text and select words like you would in any viewer, or type a term into Find and mark every match on every page. Presets catch emails, phone numbers, card numbers, Social Security numbers and IBANs.
- Apply the redactions. Every marked page is rebuilt with the content underneath each mark gone. Metadata, comments, form fields and attachments are cleaned out too, unless you keep them.
- Verify and download. The result opens in the same viewer. Try selecting text over a mark: there is nothing to select. Then download.
Covering is not removing
The most common redaction failure is a black rectangle drawn in an ordinary PDF editor. A PDF page is a list of drawing instructions, and the rectangle is simply one more instruction added to the end of the list. The text it hides is still in the file. Select-all and copy pulls it straight out. Court filings, government releases and newspaper scoops have all leaked this way.
A rectangle on top
The words sit under the box, selectable and copyable. Any script that reads the file gets them unchanged.
Rebuilt without it
The page is re-rendered with the mark painted in, and the original content stream, fonts and images for that page are dropped.
PDFslime takes the second route. When you apply, each marked page is rendered to a high resolution image, the marks are painted directly into that image, and the page is rebuilt from it. The original text, vector drawing, embedded images and annotations for that page are removed. Pages without marks are left exactly as they were.
Flattening a page normally costs you the ability to search it. To avoid that, the tool writes an invisible text layer back onto each rebuilt page, using only the text runs that did not touch a mark. Any run that overlapped a mark, even by a pixel, is dropped in full. The remaining text still searches, selects and copies as normal, and the redacted words appear nowhere in the file. You can switch this layer off in Output and clean-up if you want a pure image page.
Three ways to mark content
Area
Drag a box over anything on the page: a paragraph, a photograph, a signature, a handwritten note on a scan. Move it, resize it from the corners, or mark an entire page or page range in one go. This is the tool for scans and images, where there is no text to select.
Text
Select words exactly as you would in a PDF viewer, double-click a word or drag across a sentence, and the mark snaps to the glyphs. Multi-line selections become one mark per line so nothing between the lines is lost.
Find
Type a name, a case number or a phrase and every match across the document is listed with a snippet of context. Mark them one at a time or all at once. Match case and whole word options keep short terms precise.
Presets and patterns
One click finds every email address, phone number, payment card number, US Social Security number or IBAN in the document. Card numbers are checked against the Luhn formula so serial numbers and totals do not get caught. A regular expression option covers anything else, such as invoice or file reference formats.
What the clean-up removes
Redacting the page is only half the job. A PDF carries information in places you never see, and any of it can undo a redaction. Applying redactions therefore also cleans the file, with a switch for each group in case you need to keep one.
- Document metadata. Title, author, subject, keywords, the creating and producing applications, both dates, and the XMP metadata stream. Author fields in particular leak real names.
- Comments, links and form fields on every page. A sticky note can quote the sentence you just removed, and a form field stores its value separately from the page. Form values are flattened into the page first so nothing you can see disappears, then the fields and annotations are dropped.
- Attachments, bookmarks and scripts. An embedded file can be a full copy of the original document. Bookmark titles often repeat headings. Document-level scripts and open actions go too.
- Structure tags. Tagged PDFs keep an accessibility tree that can hold alternate text and copies of page content. It is always removed when redactions are applied, because after a page is rebuilt it would refer to content that no longer exists.
Getting redaction right
Search before you draw. A name or account number rarely appears once. Use Find to list every occurrence, including the ones in headers, footers and footnotes you would have scrolled past.
Give area marks a margin. When drawing over text or a scan, make the box a little larger than the content so no stray pixel of a descender survives at the edge. Text and Find marks add their own padding.
Leave the clean-up on. The only reason to keep metadata or comments in a redacted file is a specific requirement to do so. If you are not sure, remove them.
Use the verification view. After applying, the result opens in the same viewer. Switch to the Text tool and drag across a black box. If nothing selects, the content is gone. It takes five seconds and it is the check that catches mistakes.
Keep the original somewhere safe. Redaction cannot be undone once applied. That is the point. Store the unredacted copy where it belongs and share only the output.
Why the browser is the right place for this
Redaction exists to keep secrets. Uploading a document full of secrets to a third party so it can hide them for you is a contradiction, and once the original has left your device you have no way to know how long it is kept or who can read it. Several popular online redaction services keep files for an hour or more after processing.
PDFslime never has a copy. PDF.js, the same engine Firefox uses, renders the pages and extracts the text for search on your machine. The marks are painted in your browser, and pdf-lib rebuilds the file locally. The finished PDF goes straight from memory to your downloads folder. You can load the page, disconnect from the network, and the tool still works.
Common questions
What actually happens to the content, and what to expect from the output.
01Is the redacted content really gone?
Yes. Every page with a mark on it is rebuilt from a fresh image of that page with the marked areas painted over. The original text, vector drawing and image data for that page are removed from the file, so there is nothing underneath to select, copy or recover. The result opens in the viewer straight away so you can check for yourself.
02Does the rest of the page stay searchable?
Yes, by default. After a page is flattened, the tool writes an invisible text layer back for every piece of text that did not touch a mark, so the page still searches and copies normally. Any text run that touched a mark is left out entirely. Turn this off in Output and clean-up if you prefer a pure image page.
03Can it find every occurrence of a name or number for me?
Yes. Type a word or phrase into Find and it lists every match across the whole document with a snippet of context. Mark them one at a time or all at once. Presets look for email addresses, phone numbers, card numbers, US Social Security numbers and IBANs, and there is a regular expression option for anything else.
04What does the clean-up step remove?
Three groups, each with its own switch. Document metadata: title, author, subject, keywords, creator, producer, dates and the XMP stream. Comments, links and form fields on every page, with form values flattened into the page first so nothing visible disappears. Attachments, bookmarks and document scripts. Document structure tags are always removed on redaction, because they can carry copies of page text.
05Is this safer than drawing a black rectangle in a PDF editor?
Much safer. A rectangle added in most editors sits on top of the text as one more drawing instruction. The words are still in the file and can be selected, copied or read by any program that parses PDFs. This tool removes the content and rebuilds the page.
06Can I redact a scanned PDF?
Yes. Draw over the area you want gone. A scan has no text layer to search unless it has been through OCR, so Find and Text selection will only work on scans that already have one. Area marks work on anything.
07Can I redact a password-protected PDF?
Not directly. Remove the password with the Unlock tool first, then redact the unlocked copy.
08Are my files uploaded anywhere?
No. Rendering, searching and rebuilding all happen in your browser. The document never leaves your device, which is the only sensible arrangement for a file you are redacting.