Redact PDF

What this tool does

PDF Redact lets you click and drag black rectangles over any part of any page of a PDF — names, addresses, account numbers, salary figures, anything you don't want a recipient to see. You can stack as many rectangles as you want, across as many pages as you want, then click Apply and download a redacted file. The rendering uses pdfjs-dist, the stamping uses pdf-lib, and the whole operation runs in your browser. The file never uploads.

The output is fine for most everyday use — sharing a screenshot of a bank statement with a friend, sending a redacted invoice to a vendor, posting a contract excerpt on a forum. But there is one important caveat that the rest of this article exists to explain.

Read this first: visual redaction vs. legal-grade redaction

This is the most important thing you'll read about any browser-based PDF redactor, ours included. When you draw a black rectangle on top of text in a PDF, the rectangle is a new object drawn on top of the page. The text underneath is still there, in the PDF's content stream, untouched. Anyone with a copy of Adobe Acrobat, a free PDF text extractor, or a hex editor can pull the original text right back out.

iLovePDF and SmallPDF both have "redact PDF" features. Read their docs carefully — most of them describe exactly the visual-only redaction we're describing here, and many users never realize the underlying text survives. We say it loudly on the page and we'll say it again here: a black rectangle on top of text is not the same as removing the text.

The two-step legal-grade workflow

If you need redaction that survives a hex-editor or a text extractor, do this:

1. Step 1 — Redact here. Drop your PDF, draw black rectangles over the sensitive content on every page that needs it, click Apply, download. Output: filing-2026-05.pdf becomes filing-2026-05-redacted.pdf.
2. Step 2 — Rasterize via Compress PDF. Open our Compress PDF tool, drop the redacted file, pick any compression preset (Balanced is fine, Aggressive if you want a smaller file), download. The compressor renders every page to an image and rebuilds the PDF from those images. The black rectangles are now pixels in a JPEG layer — there is no underlying text anywhere in the document.

The tradeoff: after step 2, the PDF has no selectable text anywhere. Copy-paste won't work, search won't work, screen readers will see nothing (you'd need to run OCR to restore accessibility). That is the cost of real redaction — the people who designed Adobe Acrobat Pro's commercial Redact tool make the same tradeoff implicitly, because once content is gone, it's gone.

For everyday non-legal redaction, skip step 2. Visual redaction is enough when the recipient isn't motivated to attack the file. For everything that involves a lawyer, a regulator, or a journalist, do both steps.

Security levels: a comparison

To make the tradeoffs concrete:

The two-step workflow is essentially the free, browser-only equivalent of Acrobat Pro's Redact tool. You give up searchable text in exchange for not paying $20 a month and not uploading your sensitive documents to a third party.

How to redact, step by step

1. Drop or pick your PDF. Up to 100 MB. The first page renders to a canvas at 1.5x scale (so it looks sharp on retina displays).
2. Click and drag on the page to draw a black rectangle. Drag from the top-left of the area you want to cover to the bottom-right. The rectangle appears as soon as you release the mouse.
3. Draw as many rectangles as the page needs. Each one shows up in a list under the page with a remove button — click "×" on any rectangle to undo just that one. "Clear this page" removes all rectangles on the current page; "Clear all" wipes every page.
4. Use Prev / Next to move to other pages and draw more rectangles. The redaction count above the page navigator shows the running total across all pages.
5. When everything's covered, click "Apply N redactions." Every rectangle stamps onto the underlying PDF in one pass. Click Download.
6. For legal-grade output, take the downloaded file straight to Compress PDF.

The math under the hood: your mouse coordinates land on the canvas at 1.5x scale, then the tool translates them to PDF point coordinates (origin bottom-left in PDF; top-left in canvas, so Y is flipped), and pdf-lib draws the rectangle at the corresponding PDF coordinate. The rectangle ends up exactly where you drew it, even when the canvas displays at a different visual size than its intrinsic size.

A worked example

You're sharing a redacted version of an internal compensation report. The file is comp-report-q1-2026.pdf, 8 pages, and you need to black out every employee's salary column before sending it to a contractor who's analyzing role-level patterns but isn't allowed to see individual numbers.

You drop the file in. Page 1 renders. You can see a table with names in column A and dollar amounts in column D. You click at the top-left of the column-D header and drag down to the bottom of the last row on the page — one tall rectangle covers the entire salary column. The redaction list under the page shows "#1 ×". You click Next; page 2 has the same layout. Same drag, same rectangle. Repeat through page 7 (page 8 is appendix text with no compensation data). At the top of the navigator, the count reads "7 redactions across 7 pages."

You click Apply, download comp-report-q1-2026-redacted.pdf. This is the casual version — fine if you trust the contractor and the file won't be picked apart. To produce the legal-grade version, you upload that file to Compress PDF, pick Balanced compression, download comp-report-q1-2026-redacted-compressed.pdf. That second file is a stack of flat JPEG pages with the black rectangles baked into pixels. You ship that one.

Total time: about three minutes for the redaction itself, plus another thirty seconds for the rasterize step. Compare to running this through Adobe Acrobat Pro, which produces a similar result but requires the license, the desktop install, and (depending on settings) a fair bit of UI navigation through Tools → Redact → Mark for Redaction → Apply.

Why no upload, and what that buys you

iLovePDF, SmallPDF, and most other browser-based PDF redactors upload your file to their servers, process it, and give you a download link. For a redaction tool specifically, this is a strange design choice — you are sending the unredacted version of a sensitive document to a third-party server in order to produce a redacted version. Their privacy policies generally promise to delete the file within an hour or two. The threat model that takes seriously is "I trust this vendor to honor their policy." The threat model that doesn't take seriously is "the unredacted file existed on a server I don't control, even briefly."

PDF Redact runs entirely in your browser. The rendering uses pdfjs-dist (Mozilla's PDF renderer, also what Firefox ships with); the stamping uses pdf-lib. Open the network tab, do the whole flow, and watch zero outbound requests. There's no privacy policy to honor because there's no server in the path.

Edge cases worth knowing about

A few things to watch for:

• Color isn't black. The rectangles are pure black (#000000). If you need a different color for stylistic reasons, you can't pick one in v1. We chose pure black because it's the universal redaction color and any other choice would just confuse people.
• Rectangles, not freeform shapes. If you need to redact an irregular shape (say, a hand-drawn signature scrawled diagonally), draw multiple overlapping rectangles. Or rotate-redact-rotate using Rotate PDF if the content sits at an awkward angle.
• Form fields aren't auto-flattened. If your PDF has interactive form fields, the redaction rectangles cover them visually but the field data still lives in the form. Run the redacted PDF through Flatten PDF Form first, then redact, if you want clean output.
• Images need rasterizing. If you draw a rectangle over an embedded image (say a photo of an ID card), the image survives as image data in the content stream just like text does. Step 2 of the legal-grade workflow handles this — rasterizing flattens the image and the rectangle into a single new JPEG.

Related PDF tools

• Compress PDF — the second half of the legal-grade redaction workflow. Rasterizes every page so redaction rectangles become unrecoverable pixels.
• PDF Watermark — for adding a CONFIDENTIAL stamp to a document before or after redaction.
• Sign PDF — draw a signature and stamp it onto a page. Same canvas-based approach.
• Split PDF — useful when you only need to share a few pages and want to delete the rest entirely (the most secure form of redaction: don't include the data at all).
• PDF Merger — combine the redacted version with a cover letter or context page.

Frequently asked questions

Why isn't a black rectangle real redaction?

Because the rectangle is a new shape drawn on top of the existing page. The original text or image is still there in the PDF's content stream — anyone with a PDF text extractor, Adobe Acrobat, or a hex editor can pull it back out. For real redaction, either use Adobe Acrobat Pro's Redact tool (which removes content from the stream) or use our two-step: redact here, then rasterize via Compress PDF. After rasterizing, the page is a flat image and the rectangles are unrecoverable pixels.

How do I do the legal-grade two-step?

Step 1: redact here as normal — draw rectangles, apply, download. Step 2: open Compress PDF, drop the redacted file, pick any preset (Balanced is fine), download. The compressor rasterizes every page, so each page becomes a flat JPEG with the rectangles baked into pixels. Trade-off: the output has no selectable text anywhere. That's the cost of real security.

Is my PDF really not uploaded?

Correct. Page rendering uses pdfjs-dist in your browser; rectangle stamping uses pdf-lib. Zero outbound requests during the redact or apply steps. Check the network tab.

Can I redact on multiple pages in one session?

Yes. Use Prev/Next to move between pages — each page has its own list of rectangles. The total count is shown above the page navigator. One click of Apply stamps every rectangle across every page in one pass.

Can I undo a rectangle?

Yes. Every rectangle has a "× " button — click it to remove just that rectangle. "Clear this page" removes all rectangles on the current page. "Clear all" wipes every page. The rectangles aren't baked into the PDF until you click Apply, so corrections are free.

Does it work on a touchscreen?

Yes — pointer events cover finger, stylus, and mouse on the same code path. A mouse or stylus gives more precise control for tight redaction boundaries, but a finger on a tablet is workable for big rectangles.

What about password-protected PDFs?

pdf-lib refuses to open encrypted PDFs. Unlock the file in a desktop reader first, then bring the unlocked version here.