pdfviz

Redact a PDF

or drop one here
Choose a PDF first.

Mark private content by box, by text block, by repeated-text search, or with reviewable in-tab suggestions. Suggestions can miss private information, so you still review every page. Marked pages are rebuilt from a fresh render, and the result is verified before the download appears. The file stays in this browser tab.

How to redact a PDF online

  1. Choose a PDF

    Press Choose a PDF, or drop one on the page. The file stays in this browser tab and is never uploaded.

  2. Mark private content

    Draw a box, select a text block, find repeated text, or ask the in-tab detector to suggest people, street addresses, and patterns such as email addresses, phone numbers, card numbers and ID numbers. Review every suggestion and every page before proceeding.

  3. Choose what remains selectable

    Keep unmarked text selectable, or flatten every marked page for the strictest removal. Flattened pages lose their text layer so no one can copy what was there.

  4. Read the verification, then download

    The result is reopened. Marked regions, hidden text, raw PDF data and private metadata must all pass before the download appears.

Why a black rectangle is not a redaction

People searching for a PDF redactor online usually need one of two things: content removal or a cover. This PDF redaction tool is free and runs in this tab. Its job is to remove marked content rather than hide it. The common mistake is to draw a filled box over a name in a PDF editor and send the file. The box is one more drawing instruction added on top of the page. The name is still in the content stream underneath it, so anyone can select the text through the box, copy it out, or read it in a text editor. The page looks clean, but the information is still there.

This tool does something different. It removes the text from the content stream itself, then rebuilds the page from a fresh render. The reopened output must pass the region, text and raw-data checks before the download appears. After redaction, run the redaction checker to confirm that no selectable text remains under your boxes. If you also need to strip the author name, software or edit dates the file carries, follow up with remove PDF metadata or a full privacy scan.

When you need to redact a PDF

Legal documents and court filings. Remove names, addresses, account numbers and medical details before filing or sharing. Where the filing requires real redaction, a black cover box is not enough.

Government and public records requests. If you are publishing records released under a public-records request, check what your request and policy require, then redact personal information before the files go online.

Healthcare records. Remove patient identifiers from documents shared with researchers, insurers or second-opinion providers.

Business contracts. Redact pricing terms, partner names or proprietary clauses before sharing a contract template externally.

Academic and research files. Remove reviewer names, grant details or participant data from documents shared for collaboration.

HR and employee records. Redact social security numbers, home addresses and medical details from documents shared with managers or external auditors.

Real estate and financial disclosures. Remove account numbers and personal identifiers from closing documents before sharing with multiple parties.

How permanent PDF redaction works

A PDF page is a sequence of drawing instructions: show this text here, draw that image there. When you draw a black box over text in a standard editor, the editor adds one more instruction on top. The original text instruction is still in the stream, so anyone who reads the raw content or selects through the box can recover it.

This tool takes a different path. When you mark a region, it identifies the text and image instructions inside that region, removes them from the content stream, and rebuilds the page from a fresh render. The marked words are gone from the file, not just hidden. The rebuilt page is then checked for residual text, hidden layers and raw data before the download appears.

When you choose to flatten marked pages, the tool goes further: it renders the page to an image and drops the text layer entirely. No one can select, search or copy text from that page because the text instructions no longer exist. Use this for the strictest removal, and keep text selectable when the unmarked content needs to remain searchable.

Questions about redacting PDFs

Is my PDF uploaded?
No. The PDF, your marks, the redaction operation and the verification all run in this browser tab. No part of the file is sent to a server.
Does redaction permanently remove the text?
Yes. The marked content is removed from the PDF content stream and the page is rebuilt from a fresh render. The words are gone, not just covered.
Can I redact images?
Yes. Draw a box over any image area and the tool removes it from the rebuilt page.
What does the in-tab detector find?
It suggests people names, street addresses, and patterns such as email addresses, phone numbers, card numbers and ID numbers. Suggestions are a starting point and you review every one before applying.
Should I flatten marked pages?
Flattening a marked page removes its text layer entirely, so no one can copy or search for residual text. Use it for the strictest removal. Keep text selectable when you want the unmarked content to remain searchable.
How do I verify the redaction worked?
The tool reopens the output and checks marked regions, hidden text, raw PDF data and private metadata. For an independent check, run the result through the redaction checker.
Does redaction remove metadata too?
Redaction targets page content. To strip metadata, use remove PDF metadata or privacy scan after redacting.
Is there a page or file limit?
There is no server-side cap: processing happens in your browser. The in-tab detector suggests across up to 250 pages and 2,000 suggestions; beyond that, use find-text or draw boxes. Very large files are still bound by your browser tab's memory.

The other pdfviz tools

All of them run in a browser tab like this one, and none of them needs an account.