From file picker to memory
The browser File API grants this tab access only to files you explicitly choose. JavaScript reads those bytes into browser memory; choosing a file does not send it to the website host.
Processing and download
Canvas encodes images, JSZip reads and writes DOCX/ZIP packages, and PDF.js interprets PDF data. The browser creates a Blob URL for the result and starts a local download. Blob URLs are revoked after use.
What the network is used for
The network fetches the static HTML, CSS, JavaScript and bundled worker files that make up the site. Core libraries are served from the same site, not from a processing API or third-party CDN.
Verify it yourself
Open browser developer tools, select the Network panel, clear the list, then process a test file. You should see no request whose body contains the selected file. The large processing work appears in browser CPU and memory instead.