Redact a PDF (remove text for good)

Cover names, numbers and other private details in a PDF so they are removed from the file, not just hidden under a black box.

Mark what to cover

LOCAL & PRIVATE

Ready when you are.

Choose a PDF, or try the sample.

What is removed, and what is not

Removed for good

On every page you mark: all of its text, fonts, images, drawings and hidden layers. The page is replaced by a picture, so the covered area is black pixels and nothing is stored underneath.
Document properties: title, author, subject, keywords, creator and dates, plus hidden XML metadata.
Bookmarks, attached files, form fields and their values, comments, links and scripts, on all pages, because any of them can point at or repeat what you covered.

Not removed, or not checked

Pages you did not mark are copied as they are, with their text. If a name appears on page 3, mark page 3 too: the search does this for you.
Words in pictures or scans are not searched, and a word split in a way the PDF does not record as text may be missed. Look at every marked page before you trust the result.
Marked pages are no longer searchable or selectable, and they are larger files. The picture shows what is on the page that was not covered, including anything you did not mean to leave in.
The tool cannot see your original: keep it, and check what you share. Redaction cannot be undone.

Before the file is offered, the tool checks that none of the original page objects of the marked pages (their content, fonts and images) are in the new file, then reads the new file back with PDF.js and reports how much text it finds on those pages and whether your search words appear. Those checks are shown with the result.

The PDF is read and changed in this browser with PDF.js, which is bundled with this site. Nothing is uploaded.

A clear answer. Here’s how.

Each page you mark is drawn by PDF.js (Mozilla’s open-source PDF renderer, bundled with this site) at the resolution you choose, the marked areas are painted solid black in the picture, and that picture becomes the whole new page. The original text, fonts, images and drawing instructions of that page are not copied, so nothing underneath the black boxes can be selected, searched or recovered, and the page can no longer be searched either. You mark areas by dragging on the page or by typing numbers, or find every place a word or phrase occurs and cover it; text that is only a picture (a scan) cannot be searched and must be covered by hand. Pages you did not mark are copied unchanged, text included. The new file has no title, author or other document properties, no bookmarks, attachments or form fields, and no links or comments, because those can point at the pages that were redacted. Before it offers the file the tool compares every stored object (page content, fonts, images) of the redacted pages with the new file and refuses to produce it if any is still there, then reads the new file back with PDF.js and reports how much text it finds on the redacted pages and whether your search terms appear. Redaction is permanent: keep the original. Up to 30 MB, 500 pages and 100 redacted pages. Files are read and changed in this browser only and are never uploaded.

Appearance