Redact PDF
Cover sensitive PDF content with black boxes, using the pointer or keyboard. Download a new PDF rebuilt from PNG pages, or save the current page as PNG. The boxes are painted into the pixels and the original text layers and annotations are omitted. Processing stays in your browser; review every page and fully cover each sensitive area before sharing.
Drop a PDF here
or drop anywhere on this page · Choose a file · 100 MB
Accepted: PDF
How does Redact PDF work?
A PDF shape that covers text can leave the original characters in the file. This tool instead renders each page to pixels, paints opaque black boxes, and builds a new PDF from PNG page images. Original text layers and annotations are not copied.
Draw a box with a mouse, pen or finger. Drag an existing box to move it and use its corner to resize it. The Add a box button also works with a keyboard: arrow keys move, Shift plus arrow keys resize, and Delete removes a focused box.
Each page keeps its original size in PDF points. Export uses up to 2.2 pixels per point, with canvas limits for large pages, and lossless PNG encoding to avoid JPEG artifacts along redaction edges. The whole export fails if a page cannot be rendered.
Review every page and cover the complete sensitive area with a margin. Text outside boxes remains visible and can be read by OCR even though it is no longer selectable. The output can be larger and less accessible than the source; keep the original privately if you need a searchable copy.
Source document
invoice.pdf — 2 pages, 180 KB
page 1: drag over the account line
page 2: drag over the address block
output: every page rendered at ~160 DPI,
boxes drawn in black,
rebuilt as imagesFlattened PDF
invoice-redacted.pdf — 1.1 MB text layer: none selectable characters: 0
What options and edge cases does Redact PDF support?
| Parameter | Type | Default | Behaviour & edge cases |
|---|---|---|---|
| Draw | drag on the page | — | Draw with a pointer or use Add a box and the keyboard. Coordinates stay relative to the page. |
| Box size | normalised | — | Move and resize boxes within the page; tiny accidental drags are discarded. |
| Render scale | pixels per point | 2 | Export uses up to 2.2 pixels per point, limited to 8 million pixels and 8192 per dimension. |
| Output | flattened PDF | — | Every page is rebuilt from a lossless PNG at its original point dimensions. Original text layers and annotations are not copied. |
| Single page | PNG | — | Download the currently rendered page with the boxes painted into PNG pixels. |
| Encrypted PDFs | password | declined | Password-protected documents must be unlocked before rendering. |
Frequently asked questions
Can a recipient remove the black boxes?
The exported PDF is built from PNG pages with the boxes painted into the pixels, without the original text layer or annotations. Cover the full sensitive area and review every page. Information outside boxes remains visible and can be read by OCR.
Why is the output larger?
Text and vector pages are replaced by raster images. Lossless PNG avoids JPEG artifacts at the box edges but can increase file size.
Why is the text no longer selectable?
The output contains page images rather than the original text layer. Visible text outside the boxes can still be read or recognized with OCR. Keep the source privately if you need a searchable copy.
Is the document uploaded?
No. pdf.js renders the document in the browser, the boxes are painted on local canvases, and a local worker assembles the new PDF. Document contents are not uploaded.
Can I use a scanned PDF?
Yes. Scanned pages are rendered and redacted in the same way. Resolution and colors can change during rendering, so review the output.
Can I edit boxes with a keyboard?
Yes. Choose Add a box, then use arrow keys to move it, Shift plus arrow keys to resize it, and Delete or Backspace to remove it. Password-protected PDFs must be unlocked before rendering.