FAQ

Really, and you can check without taking my word for it: put the machine in airplane mode, load a document, and watch everything keep working. What downloads the first time is the program; the file never goes up. The code is published so anyone can audit it.

Of your files, nothing: they are never uploaded anywhere, so there is nowhere to keep them. There are no accounts and no sign-up. The only thing we count is page visits, with Vercel's counter: no cookies, no following you across sites, and no idea who you are. That counter never sees what you drop here. And your browser keeps whether you prefer light or dark, and which language you read it in.

That is what it is for. Since the file never leaves your machine, there is no third party you are handing data to and no server of ours where anything is stored. Responsibility for the document stays with you, exactly as when you open it in your usual editor.

What you already have would keep working: these are pages that run in your browser, not a service that gets switched off. The code is published, and anyone running the installed version inside their network depends on us for nothing.

By analysing the file. Drop anything, get the list of everything it says about you, and remove it in one go. The individual tools are there for when you already know what you are looking for.

It deletes. The text disappears from the document; no black rectangle is painted over it. When it finishes, the file is read back to check that nothing copyable is left underneath. If the redaction falls on an image, its pixels are erased too and the image is packed again.

Five per batch. Since all the work happens on your computer, past that point the tab starts to suffer. Plainly: that limit lives in your browser and anyone who reads the code can step over it. It is a line, not a wall. I would rather you heard it from me than found out later and assumed I was trying to fool you.

You can redact them by hand, because they are images and the pixels are erased just the same. What is not there yet is OCR to find data inside a scan: that is on the Pro plan.

Nothing, and it carries no watermarks and nothing held back. What exists is complete. What is coming will be paid for what it costs to sustain, never for putting a door in front of something that already worked.

Because a cloud API means uploading the document, which is exactly what this tool avoids. What you can do is take the engine with you: installed in your network, inside your product, or called from your own processes, always on your machine.

It helps, but it does not sign for you. Stripping the metadata from a filing before you send it is good practice and this does it in one click. What we are preparing so you can prove it is the signed certificate: the file hash, the date and the detail of what was removed. Only the hash reaches the server, never the document.