Hume AI is testing Octave 2 Multilingual internally, a text-to-speech model with low-latency voice synthesis across 10+ languages for real-time audio.

What's new? Octave 2 is a TTS model with sub200ms response, 11 languages, voice conversion and phoneme editing soon; EVI 4 mini powers speech-to-speech tasks with external LLM;

What's new? TADA is a TTS system with a text-acoustic dual alignment tokenization now open source; it produces real-time speech five times fast and supports 700 sec audio;