LTX voice: text to speech with voice design or a saved voice_id, one warm pool (merged ltx-voice + lock).
This model doesn't have a readme.