About OnsitePDF

How it works

Every tool processes your file entirely inside your browser, in a background Web Worker. The page downloads the code it needs, then your PDF is read, changed and saved on your own device. There is no upload step and no server-side processing.

The work is done by open-source libraries that run locally: pdf-lib for editing and assembling pages, PDF.js for rendering previews, qpdf-wasm (WebAssembly) for compression and encryption, and Typst (typst.ts) for generating documents.

How to verify nothing uploads

You do not have to take our word for it. You can check in under a minute.

1. Open your browser's developer tools and switch to the Network tab.
2. Open any tool, add a file and run it.
3. Watch the Network tab while the tool processes. No request carries your file data.

The first time you use a tool you may see the page download its own code (JavaScript and WebAssembly files) from this site. Those are downloads to your device, not uploads: none of them sends a request body, and none goes to another site.

Two safeguards back this up. First, tool pages are served with a locked Content Security Policy: its connect-src lists exact origins only, with no wildcards, and all third-party analytics, ad and sign-in origins are removed from it on tool pages. The browser itself refuses connections to any origin not on that list. The one third-party origin it keeps is our error-reporting service (Sentry), so we can hear about crashes. Error reports are scrubbed before they are sent: file contents, file names and extracted text are stripped out. Second, our automated test suite runs a network-guard test on every change. It processes a file in each tool and fails if any request leaves the site or carries a request body, including a request to this site itself.