ton4eg

Way more entertaining than I would expect! What TTS and ASR models do you use? What sort of latency do you get?

Do they still not have a native backup mechanism?

It's an early prototype, stay tuned!