Drop PDFs or images here
PDF, PNG, JPG or WebP. Files never leave your device.
No file is selected. Click a file in the list below to edit it, or tick several to edit them together.
Drop PDFs or images here
PDF, PNG, JPG or WebP. Files never leave your device.
No file is selected. Click a file in the list below to edit it, or tick several to edit them together.
Click a file to edit it. Ctrl/Cmd-click or tick boxes to pick several, Shift-click for a range. Several selected files are shown one after another in the viewer.
Add one or more PDFs or images (PNG, JPG or WebP) with Add Files, or drop them on the page. Each file appears as a preview in the Your Files list under the tool. Click a file to edit it, or tick several (Ctrl/Cmd-click, or Shift-click for a range) to edit them together: their pages are shown one after another, so you scroll down from one file into the next. Use Delete Selected, Delete All or a file's trash button to remove files, and Download Selected or a file's download button to save redacted copies. In Draw mode, click and drag on the page to place a redaction box. Switch to Select mode to move a box, drag its corner handles to resize it, or press Delete to remove it. Use Ctrl+Z to undo.
Type a word or phrase and choose Find & Mark All to place a box on every match across all pages of every selected PDF. The Sensitive Data Finder looks for emails, phone numbers, SSN-style numbers, card numbers, dates and URLs. It lists what it found so you can untick false matches before marking them.
Choose Export Redacted File (or a download button in the file list). Any pending boxes are applied, a new file is downloaded for each file you chose, one after another (your browser may ask to allow several downloads), and the tool re-reads it and reports whether the check passed. For PDFs, pick one of three output types first:
With a Regular PDF, each page's content is rewritten: the characters under a box are removed from the page's text, and a solid box is drawn on top, so the rest of the page stays real text in its original position. With a Flattened or Searchable PDF, each page is drawn to a canvas, the boxes are painted into the pixels, and the new PDF is built from those page images, so nothing can be recovered from them. For images, re-saving from a canvas also drops the original EXIF data such as camera details and GPS location. Annotations, form fields, bookmarks, attachments and document metadata are never carried over.
After export, the tool re-opens the new file and samples each box to confirm it is filled. For a flattened PDF it also confirms the file contains no text and that none of the redacted strings it knows about appear. For a searchable or regular PDF it confirms no text sits under any box area. A warning means you should review the output carefully.
Drawing a black rectangle over text in a PDF usually hides nothing. The text is still stored in the file underneath, and anyone can select, copy or search it. The NSA's redaction guidance lists covering text with black highlights and rectangles as a common mistake, and says the way to avoid exposure is to make sure sensitive information is "not just visually hidden." Wikipedia's article on redaction describes the same traps: placing an image layer over text without removing the text, and setting the background color to match the text color. [NSA issues redacting guidelines, Nextgov] [Redaction, Wikipedia]
A December 2009 report described how the TSA had posted its operating manual on a procurement website earlier that year with only visual redactions. Anyone who copied the text and pasted it into another program could read the covered sections, which included settings for X-ray machines and explosives detectors. In January 2019, lawyers for Paul Manafort filed a PDF with black bars over some text. The text underneath could be copied and pasted, and it included an allegation that he had shared polling information with Konstantin Kilimnik. [In wake of TSA breach, a refresher on redacting PDFs, Nextgov] [Manafort Lawyers Accidentally Reveal Trump's Campaign Chair Shared Polling Data With Alleged Russian Intel Officer, Fortune]
Files carry more than what is visible on the page. The NSA lists hidden information in office files such as Word documents as something editors often overlook, and the TSA report names leftover metadata and document properties as well. Photos have the same problem: Exif metadata from a GPS-enabled camera can record where and when a picture was taken along with the device's unique ID, often without the user's knowledge, and a cropped photo can still contain an embedded thumbnail of the original image. [NSA issues redacting guidelines, Nextgov] [Exif, Wikipedia]
This tool runs entirely in your browser, so there is no upload to intercept or scrape. That matters for the documents people redact most, such as contracts, medical records, IDs and legal filings. A March 2025 report on a warning from the FBI's Denver office said scammers imitate popular free online converters for PDF, DOC and JPG files, and that such sites can install malware or scrape uploaded files for Social Security numbers, passwords, bank account data and cryptocurrency information. [FBI warns about online file converter sites which install malware, steal private information, CBS Colorado]
"Redact" comes from the Latin redactus, the past participle of redigere, meaning "to drive back, force back; bring back." It entered English in the late 14th century meaning "combine in a unity," and by 1851 it had the sense of arranging or editing a text. The sense used today, to black out or remove parts of a document, is the first one Wiktionary lists. [redact, Online Etymology Dictionary] [redact, Wiktionary]
Before computers, printed documents were redacted by photocopying them and blacking out the text with a wide pen, and maps were covered with slips of white paper. Wartime mail was handled in a similar way: mail from the front was often opened and the offending parts blanked or cut out, and by February 1943 the US Office of Censorship had a staff of 14,462 people. [Redaction, Wikipedia] [Postal censorship, Wikipedia]
President Lyndon B. Johnson signed the US Freedom of Information Act on July 4, 1966, and it took effect a year later. Its nine exemptions let agencies withhold categories of sensitive information, and agencies may also redact the exempt portions of otherwise releasable documents, specifying the exemption for each redaction. [Freedom of Information Act (United States), Wikipedia]
Electronic redaction turned out to be harder than paper, because data often persists in storage, undo buffers, revision histories and metadata. In May 2005, a US military report on the death of Nicola Calipari was redacted by placing opaque blocks over the text in a PDF, and readers copied and pasted the covered text into a word processor. In May 2006, text on pages 12 to 14 of an AT&T legal brief was also incorrectly redacted and could be retrieved. [Redaction, Wikipedia]
On December 13, 2005, the NSA issued a report titled "Redacting with Confidence: How to Safely Publish Sanitized Reports Converted from Word or PDF." It recommended copying the sensitive document and editing the copy, pasting the text into a new blank document, and saving that clean version as a PDF. The report acknowledged that the method is slow, but said it gives the highest confidence that sensitive information is not hidden in the file. [NSA issues redacting guidelines, Nextgov]