Redact a PDF without uploading it

Why a black box is not a redaction

Every PDF viewer can draw a filled rectangle, and that is what most people reach for. It changes how the page looks and nothing else: the words stay in the file under the shape, and selecting the area and pasting it somewhere brings them back. Removing a page is not enough either, because a PDF keeps its content as separate objects, and deleting the page from the page list can leave those objects sitting in the file.

This tool avoids both. A page with a bar on it becomes a picture of the page with the bar already part of it, and the exported document is assembled new, with nothing from that page copied across. The long version, with a demonstration you can run on your own file and an account of the earlier version of this tool that got it wrong, is on how to properly redact a PDF.

What it costs you

Turning a page into an image removes the text layer for that whole page, not just the part you covered. Every other word on it stops being selectable and searchable, screen readers can no longer read it, and the file gets larger. That is the price of the covered text genuinely being gone, and the export screen says which pages will pay it before anything is written.

Pages you do not mark keep their text exactly as it was. So the cost is per page, and it is worth marking only the pages that need it. A page that would come out very large, which is usually a scan, is saved as JPEG rather than PNG. That keeps the file much smaller, at a quality that is hard to tell apart on screen.

Finding what to remove

Marking by hand works when you know where everything is. It fails on the occurrence you did not see: the account number repeated in the header of page four, the name in a footer. Find searches every page for any word or number you type, or for a whole list of them pasted in at once, one per line. It can also look for Social Security numbers, card numbers, bank routing and account numbers, employer ID numbers, phone numbers, email addresses and dates. Numbers are checked, not just matched: a card number has to pass the same checksum the banks use, and a routing number has to pass the ABA check.

You can leave the last four digits visible, which is the formfederal court rules require for Social Security and financial account numbers in filings. Every match becomes a bar you can see and adjust before export, and nothing is applied until you say so. Find reads the text layer, so it cannot see a scanned page that has none; the panel tells you which pages those are.

Using it offline, or as an app

After your first visit, the tools keep working with the network switched off, because the browser keeps a copy of everything they need. You can also install the site as an app with its own window. In Chrome or Edge on a computer, use the install icon at the right of the address bar or the install option in the browser menu. On Android, choose Install app or Add to home screen from Chrome’s menu. On an iPhone or iPad, tap Share in Safari, then Add to Home Screen, and on a Mac, choose File, then Add to Dock in Safari.

Installed from Chrome or Edge on a computer, it also appears under Open with when you right-click a PDF in your files; the first time, the browser asks whether to allow it. The installed app is this same site, so the document is read from your disk and still never leaves it.

What it does not handle

Password-protected PDFs are refused rather than half-processed; save an unprotected copy from the app that made the file and bring that. XFA forms, produced by some older enterprise software, cannot be flattened by the library underneath, and the tool says so when it meets one. A digital signature does not survive any rewrite, so a signed document comes back unsigned; your original file is not changed.

Metadata inside a photograph embedded on a page you do not mark is not cleared, because it belongs to the picture rather than the document. A photo you open here directly is different: it is converted to a page first, which strips its location and camera details. Very large files are the honest weak spot. Everything runs in this tab, so a big scan is slow, and on a phone it can run the browser out of memory.

Common questions