Changelog

EddyLoop no longer costs anything, and there is nothing to sign in to. The account, the licence, the trial clock and the five machine limit are gone, along with every part of the app that used to check them. Download the installer, run it, and it works. Put it on as many of your own machines as you like.

The app no longer asks the network for anything on its own account. It does not check a licence, it does not look for updates, and it has never sent a word of what you say anywhere. The only request it ever makes now is downloading the speech model once, from Hugging Face, when you choose one.

If you are reading older entries below, they describe a version of EddyLoop that was sold. Everything they say about signing in, trials, licences and device limits stopped being true on this date.

EddyLoop now opens on a sign in screen the first time you run it, rather than a settings list you cannot use yet. Signing in checks your license in the same step, so the app is ready to record the moment it lets you through, and it says so on the spot if something still needs sorting out instead of leaving you to find out when a recording will not start. Clicking one of EddyLoop’s notifications now opens the app.

EddyLoop now signs in to your account from Settings. Signing in starts your free trial, and it unlocks a purchase on every machine you use, up to five. The app checks in with the site now and then and keeps working offline in between, so a flight or a dead router never stops a recording. Machines you no longer use can be freed from your account page.

Quiet stretches and background noise could put sentences into the log that nobody said. A segment now has to hear a run of speech before it opens one, and sounds that are not speech no longer arrive as text. Stopping a session no longer waits on a stalled write either, so the log closes when you ask it to.

EddyLoop listens in the background and writes what it hears to a plain text file on your machine. Nothing is uploaded. The tray icon tells you whether the microphone is live, and every line is written as you speak, so an unexpected reboot costs seconds rather than the session.