[jinto36]

Hi, I'm just wondering what the text-to-speech and speech-to-text for Japanese is based on? It works pretty well, and speech-to-text in a browser is not something I thought would be practical a few years ago. How much of that is being done client-side? Also like others I recognize some of the "voices" from Duolingo, so presumably there's some text-to-speech engine in common? The prompt/scene generator is amusing.