My regex matches the right text but the rectangles are wrong.

Pdfjs reports text positions per "item" (often a span of letters). When a match crosses item boundaries, every overlapping item gets a rect. The result tends to slightly over-redact rather than under-redact, which is the right error mode for privacy.

Does it support unicode classes (\p{L} etc.)?

JavaScript supports them with the u flag. Add "u" to the flags field.

Redact PDF by Pattern

Provide a regex; every match is rasterised over with a black box on each page. Catch SSNs, internal codes, names, anything that follows a stable pattern.

Your files never leave your browser. All processing happens on your device — nothing is uploaded.

1. Upload your file

Drop a PDF file here

or click to browse

How does the Redact PDF by Pattern work?

Extract per-page text positions with pdfjs, run a JavaScript regex over the concatenated text, and draw a black rectangle over each matching text fragment. Default pattern matches US SSNs; provide your own regex to catch internal account numbers, employee IDs, project codenames, or anything that follows a stable shape. Note: this is VISUAL redaction (black rectangles drawn on top of the original text). For TRUE permanent redaction (rasterise + black-out so the underlying text cannot be extracted), pipe the result through Permanent Redaction.

🔒 Security & Privacy

All processing happens on your device. Your files never leave your browser, never touch our servers, and are not stored anywhere. Close the tab and the file is gone.

📱 Use it on any device

Works in any modern browser — Mac, Windows, Linux, iPhone, Android, and tablets. No installation, no app to download. Just open the page and start.

⚡ Quick & easy

Drag, drop, click. Most files process in seconds, not minutes. No watermarks added to your output, no per-day limits, and the tool always shows the result before asking you to download.

🎁 Always free, no signup

Free to use, every time. No account creation, no email required, no “trial” that converts to a paid plan. We make money from ads so you don’t have to pay.

Frequently asked

Visual is sufficient if your only concern is humans reading the document. For documents going to adversaries who might extract text via copy/paste or programmatic extraction, the visual rectangles can be deleted. Use Permanent Redaction afterward for those cases.