A 1 billion parameter text-to-speech model by Hume AI, designed for emotionally expressive and natural speech synthesis.