Check it yourself

1. Watch the network traffic during a job

1. Open a tool, for example Merge PDFs, then the browser's developer tools (F12, or Ctrl/Cmd + Shift + I) and the “Network” tab.
2. Tick “Preserve log”, drop your files, run the job, download the result.
3. Look at the requests: only files from the site itself are loaded (scripts, fonts, WebAssembly engines under /pdfjs, /ocr, /qpdf, /fonts, /heif). No POST request, no upload to another domain, no request whose size would match your document.

With the default configuration, the site loads no third-party service. If the publisher enables analytics or advertising, it only loads after your consent and never has access to the content of your files.

2. Cut the connection

Open the tool's page, then switch to aeroplane mode or turn off Wi-Fi. Drop a file and run the job: it completes, because all the code needed was loaded with the page. With “Use offline”, the tool even stays available for later visits without a network (see the offline page).

3. Read the code

Processing relies on open-source libraries you can inspect: pdf.js and pdf-lib (PDF), Tesseract.js (text recognition), qpdf (encryption), libheif (HEIC photos), mammoth, exceljs and docx (Office conversions). They are served by the site itself, without a third-party CDN, and run in your browser.

The site's security policy (the Content-Security-Policy header, visible in the Network tab) forbids the code from contacting any domain other than the site itself: even a programming mistake could not send a file elsewhere without the browser blocking it.

What this does not cover

• Your browser and its extensions may themselves send data (sync, capture extensions): this does not depend on the site.
• The downloaded file is saved on your device, in the downloads folder: protecting or deleting it is up to you.
• Nothing is stored by the site, but the browser's memory keeps the document open as long as the tab is: close the tab on a shared device.