Guide

How a file is processed without uploading it

Damaged files are often private ones. Here the program comes to the file, so the file never travels.

How it works

The engine is compiled to WebAssembly, which browsers run directly. The page downloads it once, and the work then happens on your processor, in background workers so the page keeps responding.

How to check it

Open the browser's developer tools on the Network tab, then process a file. Requests fetch the program's own files, and the damaged file is not among them. You can also disconnect after the page loads.

The trade-off

Everything is held in the browser's memory: the file and the result at once. On a phone, or with very large files, that memory can run out. The page says so when it happens.

Common questions

Does the engine run on every core?

Where the browser allows it, yes. The page reports how many cores are in use.

Is anything stored in my browser?

The result is held in memory until you save it or close the page.

Sources

Documents consulted for this guide.

  1. WebAssembly
  2. MDN: SharedArrayBuffer