Text to speech

Indic Parler-TTS by default (language, speaker, emotion, pace, and pitch controls), or IndicF5 for voice cloning from a reference clip.