Comment by narrationbox

Plenty of models take in text + audio and spits out audio. It's the format of most newer generation accent conversion/voice cloning models.

What's your exact use case?