Voice cloning — Tamil & 33 other languages

Upload a short reference clip, pick a language, and the model speaks your text in that voice.
IndicF5 covers Tamil and 10 other Indian languages. Chatterbox covers English and 22 more. Both are MIT licensed, so output is usable commercially.

Language

Reference voice (5–15s, clean, one speaker)

Reference transcript — exactly what is said in the clip

Required by IndicF5. Accuracy here strongly affects the result.

Text to speak

I have the speaker's permission to clone this voice.

Cloned speech

Tips
• A clean, dry 5–15s clip beats a long noisy one.
• One speaker only, no music or background.
• For IndicF5 the transcript must match the clip word for word.