ToolJutsu
All tools
PDF Tools

Redact PDF

Permanently remove text from a PDF by rasterising redacted regions.

Redaction is permanent. The exported PDF rebuilds every redacted page as a flattened image with black rectangles painted in. The original text under the black bar is not present in the output file's content stream — text-extraction tools, copy-paste and `pdftotext` cannot recover it. Treat the output as a one-way transformation.
Processed on your device. We never see your files.

How to use Redact PDF

What this tool does, and why the distinction matters

PDF redaction is one of the easiest things to get wrong. There are years of public incidents — court filings, government reports, corporate disclosures — where a document was “redacted” by drawing black rectangles in Acrobat or PowerPoint over the sensitive text, saved as PDF, and published. The text was still there. A reader selected it, copy-pasted it out, and the redaction failed in public.

A 2023 federal court filing in the United States had a redacted PDF where the underlying text was extractable with a free command- line tool inside a minute. Variants of this story have happened every year for two decades — to law firms, banks, defence ministries, and Fortune 500 companies.

The lesson is simple: a black rectangle on top of text is not redaction. It’s a sticker.

How this tool actually redacts

The pipeline is rasterisation-based, deliberately:

  1. For each page you redact, pdf.js renders the original page to a 2x-DPI canvas.
  2. The regions you’ve drawn on the page are filled with solid black on that canvas. This paints over the rendered pixels — not over the underlying PDF text, because at this point we’re working with pixels.
  3. The canvas is exported as a PNG.
  4. pdf-lib opens the original PDF, replaces each redacted page’s content with an image-drawing operator that draws the PNG at full page size, and saves a new PDF.

The output for redacted pages is a flat image. There is no text layer, no font information, no vector content underneath the black fill — because the entire page was reduced to pixels before the black fill was applied. There is nothing left to extract.

Pages you didn’t redact pass through unchanged, with their original text layer intact.

Common use cases

  • Legal filings. Pleadings, motions, exhibits, deposition transcripts that need party names, social-security numbers, account numbers, or sealed information removed before public filing.
  • Contracts containing personal information. Salaries, addresses, bank details, identification numbers being shared with a counterparty that doesn’t need them.
  • Screenshots of dashboards. Internal analytics, customer lists, revenue figures, or production data being shared in decks, blog posts, or social media — redact before publishing.
  • FOIA-style document release. Public-records responses where some content is releasable and other content is exempt.
  • Investigative reporting source-protection. Source names, contact details, and routing information in document leaks being prepared for publication.
  • Sanitising old PDFs for archive. Removing personal data before moving files into a long-term shared archive.

What this tool doesn’t do

  • It doesn’t preserve the text layer. Redacted pages become images. Unredacted pages keep their text. Mixed-search expectations should be set accordingly.
  • It doesn’t redact metadata automatically. Authors, titles, edit history, embedded thumbnails, and XMP metadata are separate from the page content and can leak sensitive information by themselves. Use PDF Metadata Remover after redacting if metadata might be sensitive.
  • It doesn’t audit what you missed. The tool redacts the regions you drew. Identifying every instance of every sensitive string in a 200-page document is on you (or an automation built around full-text search and review).

How to use this Redact PDF tool

  1. Drop a PDF onto the dropzone. The first page renders as an interactive canvas.
  2. Draw rectangles over each region you want redacted. Each rectangle is shown in a translucent red overlay so you can see exactly what will be blacked out.
  3. Navigate to other pages with the Prev / Next buttons and draw redaction regions there too. Each page tracks its own regions.
  4. Optional: undo the last region per page, or clear all regions on the current page.
  5. Click Apply Redactions. The tool rasterises each redacted page, paints the regions black, and rebuilds the PDF.
  6. A Download button appears with the redacted PDF.

Security and limits considerations

This is the most safety-critical tool on ToolJutsu, and a few things are worth being explicit about:

  • The redaction is real. Once applied, the original content underneath the black regions is genuinely gone from the output file. You cannot recover it from the output — and neither can anyone else.
  • Always verify before sharing. Open the redacted PDF, try to select text in the redacted region (you should get nothing or get the unredacted text on the rest of the page), run pdftotext on it if you have it, and visually scan every page.
  • Keep the original. The redacted PDF is a new file. Your original PDF is untouched on your disk. If a redaction was wrong (too small, missed a region, wrong region), redact the original again — don’t try to recover content from the redacted output.
  • Metadata is separate. Author name, document title, edit history, and embedded thumbnails survive redaction unless you also strip them.

Privacy

Redaction by definition involves sensitive content, so this is worth repeating: nothing leaves your device. The PDF is rendered locally by pdf.js, redacted on a local canvas, and rebuilt by pdf-lib in this browser tab. No upload, no server-side processing, no logging of document content. The only network requests this page makes are for the JavaScript bundle when it first loads.

Compatibility notes

The redacted PDF is a standard PDF 1.7 file whose redacted pages are image-backed. Every reader displays it correctly — Adobe Acrobat, Apple Preview, the browser viewers in Chrome / Edge / Firefox / Safari, mobile readers, and print pipelines. Because the redacted pages are higher-DPI images, the output file is larger than a text-only PDF; that’s the cost of real redaction.

Frequently asked questions

How is this different from drawing a black rectangle over the text?
This is the single most important question about PDF redaction, and the answer is the difference between real redaction and the famous failure mode. A black rectangle drawn on top of a PDF is just another shape in the content stream — the original text still sits underneath it. Anyone can open the file in Acrobat, delete the rectangle, and read what was 'redacted'. Worse, pdftotext, copy-paste, search, and screen readers ignore the rectangle entirely and return the original text. This tool rasterises each redacted page — it renders the page to a canvas, paints solid black over the regions you chose, then rebuilds the PDF page from the resulting image. There is no text layer underneath to recover, because there is no text layer at all on the redacted pages.
Why does my redacted PDF get bigger?
Because rasterised pages are images, and images are larger than text. A page that was 30 KB of text and vector content might become 200–400 KB as a PNG-backed page. That's the cost of certainty — the original text genuinely no longer exists in the file, and an image of the page is what replaces it. If file size is a concern after redaction, run the output through Compress PDF to re-encode the images at lower quality.
Will the unredacted parts of the page still be searchable and selectable?
No — and this is the important honest trade-off. Once a page is rasterised, the entire page becomes a flat image. Text that wasn't redacted on that page is no longer copyable, searchable, or accessible to screen readers, because there's no text layer left at all. Pages with no redactions on them keep their full text layer; only the redacted pages become image-only. If preserving search across the document matters, redact only the pages that need it, or run the output through an OCR step to re-add a text layer to the redacted images (a separate workflow).
How does this compare to Adobe Acrobat Pro's redaction tool?
Adobe's redaction tool also produces tamper-proof output — it edits the underlying content stream to remove the text, then optionally overlays a black mark. This tool takes a different but equally definitive route: rasterise the page, paint over the regions, rebuild the page from the image. The end result for the redacted regions is comparable — the original content is genuinely gone in both cases. The Acrobat approach preserves the text layer of the rest of the page; the rasterise approach loses it. The trade-off is search vs. simplicity; the redaction itself is real either way.
Does any of this leave my device?
No, and for a redaction tool this matters more than usual. The PDF you redact is, by definition, a document containing information you don't want a wider audience to see. The entire pipeline runs in this browser tab: pdf.js renders each page to a canvas, your drawn regions are painted black on that canvas, the canvas is exported as PNG, and pdf-lib rebuilds the PDF with those PNGs as the redacted pages. No upload, no cloud rendering, no analytics on document contents. Confirm in the browser Network panel — only the initial JavaScript bundle requests are visible. After that you can disconnect and the redaction still works.

Related tools

PDF Watermark Adder

Stamp text or image watermarks onto a PDF.

PDF Stamp Adder

Add DRAFT, CONFIDENTIAL, or APPROVED stamps to a PDF.

PDF Flatten

Flatten PDF layers and form fields into static content.

Photo Censor

Pixelate, blur or black-bar sensitive regions of a screenshot before sharing.

PDF Password Protector

Encrypt a PDF with a password.

PDF Text Extractor

Pull the plain text content out of a PDF.