Surge AI - Voice Model Devlog Most of my time lately has been spent on developing the voice feature in Surge AI.The previous implementation of voice in the app was rather slow and quite unnatural-sounding at times, so I’ve been experimenting with other models and trying to improve the experience as a whole rather than merely swapping the voice.There’s still some tuning required, but I can already tell that the answers sound a lot better now.It’s this kind of feature that seems like a straightforward thing to implement until you actually try doing that.