Don’t take our word for it — check

Anyone can write “we don’t upload your files”. Here are three ways to check it yourself, from the easiest to the most thorough. All of them use things already built into your browser.

1. The aeroplane test (10 seconds, no technical knowledge)

  1. Open any tool, for example Merge PDF, and wait for it to finish loading.
  2. Turn off Wi-Fi, or switch the device to flight mode.
  3. Use the tool. Merge two PDFs, convert a photo, generate a QR code.

It works. A tool that uploaded your file could not possibly work with the connection off. That is the whole argument, and anyone can run it. (The few “online tools” — DNS and domain checks — are the exception: they need a connection by nature, and each of their pages says exactly which public service it asks.)

2. The Network tab (30 seconds)

  1. Press F12 (or I on a Mac) and select Network.
  2. Press the clear button so the list is empty, then load your file into the tool and run it.
  3. Read the list. You will see requests for the tool’s own code the first time — JavaScript, a WebAssembly module, perhaps a font — and after that, nothing.

Sort by Size and look for anything going out that is as large as your file. There is nothing to find: apart from our own small page-view counter (described below), every request is a GET that fetches one of our static files. Nothing can receive an upload — this site is a folder of static files on a rented server, with no upload handler at all.

3. Read the code (as long as you like)

The pages are plain static files. Use View source, or the Sources panel in developer tools, and read what the tab is running. The heavy lifting comes from open-source libraries you can recognise and check yourself — pdf.js, libheif, ffmpeg, Tesseract — all listed with their licences on the licences page.

What we do collect

Two things, and only two. First, the web server keeps the ordinary access log every web server keeps: which page was requested, when, from which IP address and with which browser. Second, our own small page-view counter (loreatec.jp/a/t.js) sends us the page address, the page title, the page you came from, the width of your window and how long the page stayed open. It sets no cookie, stores nothing on your device, uses no third party, and stays silent if your browser sends “Do Not Track”. Neither of the two ever sees a file name or a file’s contents — the files never reach any server, so they cannot appear anywhere. That is how we know which tools are worth improving, and it is all we know.