Redact PDF — Black Out Sensitive Information

Redact PDFs the Right Way — Without Uploading Them

Before you email a bank statement, share a contract, file a court document or post a screenshot, you often need to hide a few details: an account number, a phone number, an Aadhaar or PAN number, a customer's name. This free PDF redaction tool finds that information for you with on-device OCR, lets you check every box, and then produces a new PDF in which the sensitive parts are gone for good.

Why a black box in a PDF editor is often not redaction

Drawing a black rectangle or a black highlight in an ordinary PDF editor only adds a shape on top of the page. The original words are still in the file underneath: anyone can select and copy them, find them with search, delete the rectangle, or pull them out with a text extractor. That is how many "redacted" court filings and government reports have leaked. Metadata, bookmarks, comments, form fields and hidden OCR text layers can give away the same details.

This tool avoids all of that by flattening every page to an image. Each page is rendered to pixels, the black boxes are painted into those pixels, and a brand-new PDF is built from the images. The original text, fonts, links, form fields, annotations, attachments, scripts and metadata are not copied, so there is nothing left under a box to recover.

How to redact a PDF

1. Add your file — drop a PDF (digital or scanned, any length) or one or more photos and screenshots. Each image becomes a page.
2. Let it scan — OCR reads every page in your browser and marks emails, phone numbers, card numbers, IBANs, PAN, Aadhaar, GSTIN, SSN and NI numbers and IP addresses. Type names or project codes into Also redact these words.
3. Review — click a coloured box to keep it visible, untick a whole category, or drag across the page to black out anything else, such as a signature, photo or address.
4. Download — get a new PDF with the boxes burned in. For images you can also save each redacted picture as a PNG.

What it finds automatically

• Payment cards — 13 to 19 digits with or without spaces, confirmed with the Luhn checksum so ordinary invoice numbers are ignored.
• Aadhaar numbers, validated with the Verhoeff checksum UIDAI uses, plus PAN and GSTIN (including its check character).
• IBANs checked with the mod-97 rule, US Social Security numbers and UK National Insurance numbers.
• Phone numbers — Indian mobiles with or without +91, international, US and UK formats — and email and IP addresses.
• Optional: every date, and the values after labels such as Name:, Account No:, Policy No:, Customer ID: or DOB:.

Verify before you share

• Open the downloaded PDF and look at every page at full size — is each sensitive area completely black?
• Press Ctrl+F and search for a name or number you redacted; it must not be found.
• Try to select or copy text inside a black box — nothing should come out.
• OCR cannot read photos, signatures, stamps or handwriting: box those by hand.
• Check the file name, and keep the original somewhere safe — the redacted copy cannot be turned back.

Limitations

OCR reads printed text. It can miss handwriting, very small or faint print and text inside photos, and a misread digit can make a checksum fail — always look through the pages yourself. For digital PDFs the real text layer is checked as well, which catches most values OCR misreads. Because pages become images, the result is not editable, and its text cannot be selected unless you tick Keep text searchable; that option runs OCR again on the redacted pages, so only the text you left visible becomes searchable. Big scans make big files — run Compress PDF afterwards if needed.

Private by design

Rendering, OCR, detection and building the new PDF all happen in your browser. Your document is never uploaded to SarvKit or anyone else; only the OCR engine and language data are downloaded once and cached. Read our PDF redaction guide for the mistakes that cause leaks. Need searchable text instead of redaction? Try OCR PDF, Scan to Word or Image to Text.