Redact PDF

Remove sensitive content for real — marked pages are rebuilt as images, in your browser.

Runs in your browser. Nothing is uploaded — your input is processed on this device and discarded when you close the tab.

Your file

Drop a PDF here or click to browse. One PDF at a time, up to 50 MB. Everything is processed on this device only — no file is uploaded.

Page preview — drag a rectangle over each piece of content to remove

Previous page Next page Undo last region on this page Clear all regions

Resolution for rebuilt pages 150 DPI — smaller file, fine for screens 200 DPI — larger file, sharper for print

Redaction regions

Redact and download

There is a wrong way to redact a PDF, and it is the way most people first try: draw a black rectangle over the sensitive words. The page looks safe. It is not. The text is still sitting in the file underneath the rectangle — selectable, copyable, searchable, and recoverable by anyone who deletes the shape or simply copies the page's text into another document. Real incidents, including court filings released with readable "redacted" passages, have come from exactly this mistake. This tool redacts the other way: it removes the content itself.

The method is blunt and provable. Any page where you have drawn a redaction region is rendered to a picture at the resolution you choose (150 DPI, or 200 for print), your regions are painted solid black onto that picture, and the page is rebuilt in the output as that single full-page image. The original page — its text, its fonts, its content stream — never reaches the output file at all, so there is nothing underneath the black to recover, for anyone, with any tool. Pages where you drew nothing are copied across untouched, so their text stays real text. The output document is also built from scratch, which means the source file's title, author and other metadata are not carried into it. Everything runs in your browser; a document sensitive enough to redact is the last file that should be uploaded to a stranger's server, and here it never is.

How to use Redact PDF

1. Add your PDF. Drop the file into the box above, or click it to choose from your device — one PDF, no larger than 50 MB. Files that are password-locked or damaged are flagged by name and turned away.
2. Open the page to redact. The preview shows one page at a time; move through the document with Previous page and Next page. Only pages you mark are rebuilt; every other page is copied as real text.
3. Drag a rectangle over each piece of content to remove. Drag in any direction — the rectangle is normalised for you. Cover the whole item: a name, a number, a signature, a photograph. A rectangle that is too small to cover anything is refused.
4. Review the region list. Every rectangle is listed with its page and size, and each can be deleted individually. Undo removes the last region on the current page; Clear removes them all. Check each marked page once more before applying — redaction is permanent in the output.
5. Choose the resolution. 150 DPI is the default and reads clearly on screen. Choose 200 DPI if the rebuilt pages will be printed; the file will be larger and the text edges sharper.
6. Redact and download. Pages with regions are rebuilt as images with the regions painted black; pages without are copied untouched. Open the download and try it yourself: select the redacted area and copy — there is no text there to copy, because there is no text there at all.

Features

• True redaction by rasterisation: a marked page's original content stream is discarded entirely and replaced by a rendered image, so covered content is removed, not hidden.
• Regions are painted solid black, fully opaque, onto the rendered page before it is rebuilt as a JPEG — what is under the black does not exist in the output.
• Unmarked pages are copied at the object level with pdf-lib, keeping their text selectable and their size unchanged — only the pages that need it pay the image cost.
• Two resolutions: 150 DPI by default, 200 DPI as an option, chosen per job.
• The output is a fresh document: no title, author or other source metadata is carried over.
• Drag in any direction to draw; regions are normalised automatically, listed with page and size, and can be deleted one by one before anything is applied.
• Applying with zero regions drawn is refused, with an explanation — the tool will not hand you an ordinary copy dressed up as a redaction.
• Region coordinates are converted carefully through three systems — screen pixels, PDF points and image pixels, including the vertical flip between them — by tested helpers, so the black lands where you dragged it.

Worked example

Take a two-page A4 letter (595 × 842 points) whose first page carries a reference number at the top: the words occupy the area from 40 to 230 points across and 750 to 795 points up from the bottom-left corner. Draw one region over exactly that area on page 1 and nothing on page 2, and apply at the default 150 DPI. The tool's plan is then: rasterise page 1, copy page 2. Page 1 is rendered at 150 DPI — a 1,240 × 1,755 pixel image — the region is painted black on it, and the page is rebuilt as that image. In the output, extracting page 1's text returns nothing at all: the whole page is a picture, so the reference number is not merely covered but absent from the file. Page 2, copied untouched, still yields its full text. A pixel read from the middle of the black region is (near) (0, 0, 0) — JPEG encoding keeps it within a few points of pure black — while a pixel in an unmarked corner of the same page stays white, above 240 on every channel. Each figure in this walkthrough is asserted by the tool's automated test suite, so the example cannot quietly drift away from the code.

Common errors and limitations

• A rebuilt page becomes a picture of a page. Its text can no longer be selected, searched or copied — including the text you did not cover. That is the price of certainty on that page, and it is why unmarked pages are left as real text.
• Rebuilt pages are JPEG images, so fine text is slightly softer than the original vector text, more so at 150 DPI than at 200. For a document where a whole book page must stay pin-sharp in print, expect a larger file at 200 DPI.
• Redaction here covers rectangular regions on pages. It does not hunt through the file for patterns: attachments, embedded files, layers, comments, form-field values and document metadata beyond the fresh-document rebuild are outside its scope. Check a sensitive file's attachments and properties separately before sharing.
• The black is only where you drew it. A region that misses the last digit of an account number leaves that digit in the rendered image, visible to any reader. Draw generously and re-check each marked page in the output.
• One PDF per run, capped at 50 MB; a long document with many marked pages takes time to render page by page.
• Password-protected PDFs cannot be opened here at all.
• Redaction is permanent in the output. Keep the original somewhere safe and private; the redacted copy cannot be un-redacted.

• Drawing a black box in an ordinary editor instead. That hides text visually and leaves it fully recoverable in the file. Fix: use this tool, whose marked pages are rebuilt as images with the original content discarded.
• Drawing the region a fraction too small. The visible glyphs are covered, but a descender, a digit or the edge of a signature survives in the image. Fix: drag well past the content on every side, then zoom into the output page and look before sharing.
• Redacting one copy of the data and missing its twin. The same number often appears in a header, a footer, a summary table and a reference line. Fix: search the original for every occurrence of the sensitive value and mark each page it appears on.
• Sharing the wrong file afterwards. The original and the redacted copy look identical in a file list at a glance. Fix: download the redacted output, open it, and verify — try selecting the redacted area — before it leaves your device; keep the original out of shared folders.

Frequently asked questions

Related PDF tools

Prefer everything on one page? The original PDF Toolkit merges, splits and rearranges with one shared file list. The full suite is indexed on the PDF Tools hub.

How it works

Open the Redact PDF page and choose a PDF from your device. Everything runs in your browser — the file never leaves your device. You mark the regions that contain sensitive content, such as names, account numbers, or addresses, and the tool removes that content for real. Pages with marked regions are rebuilt as images, so the covered text is genuinely gone from the output file rather than hidden underneath a black rectangle that could be peeled away. You choose a rendering resolution of 150 or 200 DPI for the rebuilt pages: higher resolution keeps small print more legible, while lower resolution keeps the file smaller. The result downloads as a new PDF, and your original file is left untouched so you can redo the job if you mark the wrong region.

Another worked example

Jordan, a fictional HR assistant, must share a meeting report that contains one employee's private phone number and home address. Jordan opens the tool, selects the report, and marks the two regions containing those details. The tool rebuilds that page as an image with the marked areas removed, while the rest of the report stays readable. Before sending anything, Jordan opens the downloaded file, searches it for the phone number, and confirms nothing comes up. Only then does Jordan email the redacted copy to the wider team, keeping the original in a restricted folder.

More limitations to know

Redaction must be permanent removal, not a black box: this tool rebuilds marked pages as images precisely because a drawn rectangle over live text can be removed by anyone with a PDF editor. That rebuild has a cost — text on a rebuilt page becomes an image, so it can no longer be selected, searched, or copied. Redaction also only covers what you mark, so a detail repeated elsewhere in the document, in a header, or in a second copy of the file will still be visible unless you catch it. Always search the output for the sensitive strings before sending, and remember that copies you already shared cannot be recalled by redacting a new one.

More frequently asked questions

Why is my redacted page no longer searchable?

Because the page was rebuilt as an image to make removal permanent. Searchable text and removable text are the same thing here — keeping one would keep the other.

Should I choose 150 or 200 DPI?

Choose 200 DPI when the page has small print you need to stay legible, and 150 DPI when the document is mostly large text and you want a smaller file.

Can I redact only part of a page?

Yes. Mark exactly the regions that hold sensitive content, and only pages with marked regions are rebuilt; the rest of the document is left as it was.

What should I check before sending a redacted file?

Open the output, search it for every sensitive name, number, and address you meant to remove, and glance at each rebuilt page. A two-minute check is the whole point of the exercise.

Related articles