Redact PDF

Keep private information private.

Draw over sensitive areas. Export rebuilds every page as an image and removes original text, attachments and metadata. Review the result before sharing.

How to use Redact PDF

1. Open your document from your device.
2. Adjust the settings and preview the result.
3. Read text results on screen or download your file. Your original remains untouched.

Local by design

Your documents are processed on your device. No account, payment or document upload is required.

Enable JavaScript to use the document workspace.

Remove marked information from a new visual copy

Redact PDF is for information that a recipient should not receive, such as a visible identifier or private address. Unlike ordinary annotation, the export rebuilds every page as an image after covering the selected areas. It discards the original text, attachments and source objects rather than leaving a black rectangle over recoverable text.

The trade-off is important: the result loses original searchable text and interactive features. It is a visual delivery copy, not a preservation of the original document structure or a legal certification. Keep the original securely and follow any required disclosure-review procedure. The tool cannot decide every location or type of information that should be withheld.

Mark all relevant locations

1. Open the PDF and review the entire document, including appendices.
2. Draw over sensitive areas in the preview.
3. Use text snapping where available and inspect the expanded coverage.
4. Check scans manually, leaving enough coverage around letters and image details.
5. Apply redaction, download the result and independently inspect the saved PDF.

Text snapping uses intersecting text runs, which may contain several words. It can cover more than the small part you initially drew over. Review each mark and use Undo when the scope is wrong. On image-only scans, there may be no reliable text coordinates, so the drawn rectangle itself must fully cover the sensitive pixels.

Review more than the obvious name

A name can repeat in a footer, signature block, table or image. An account reference can appear inside a barcode. A job title or unique event date can identify someone indirectly. Search can help on text-based pages, but it cannot replace a full visual review, especially when scans or unusual encodings are involved.

Check rotated pages and mixed page sizes carefully. The preview’s coordinate mapping should place the mark correctly, but the exported page is the final evidence. Inspect all four edges of each covered area for exposed digits, punctuation or handwritten strokes. A tiny uncovered fragment can defeat the intended removal.

The operation rebuilds unmarked pages too, avoiding carriage of original hidden objects into the output. However, visible information on an unmarked page remains visible in its new image. Removing document metadata does not remove an author name printed in a footer. Content review and structural cleanup address different risks.

Verify the downloaded file

Open the result in another reader when practical. Check every redaction visually, count pages and inspect important unmarked content for readability. Attempt text extraction or search: the original text layer should not be present in this visual output. Also inspect attachment, comments and properties panels where available.

No single check is a universal proof. A failed search can also occur because of poor text encoding, so combine visual and structural review. If your process requires forensic inspection or a certified tool, use that required process. Do not infer legal sufficiency from a successful export message.

If you later add OCR, recognise the redacted page images, never reuse an original text layer that could reintroduce removed information. Keep delivery filenames and accompanying messages free of the details you intended to withhold. Read why black boxes fail and metadata review before distributing an important redacted copy.

Frequently asked questions

Is this different from drawing a black annotation?

Yes. The result is rebuilt from covered page images without carrying the original text objects into the new PDF.

Will the result remain searchable?

The original searchable text is removed. OCR afterwards is a separate operation that must use the redacted images.

Does snapping always identify individual words?

No. It expands to intersecting text runs, which can include multiple words. Scans may require fully manual coverage.

Is the output legally certified?

No. Follow the review and tool requirements of your particular disclosure process.

Do I still need to review every page?

Yes. The tool removes what you mark; it cannot determine every visible or indirect disclosure on your behalf.