Voice

Speech synthesis, zero-shot cloning, transcription, and dubbing — on your own models.

Synthesize

Stream (play as it generates)

Add a short, clean clip (5–10s). Zero-shot — no training step.

I confirm the speaker consented to their voice being cloned. Cloning someone’s voice without permission may be illegal.

No cloning engine installed — install Chatterbox (MIT). Mock clones for UI testing only.

Audio → text (speech recognition).

Audio → transcribe → translate → re-voice (uses the selected voice).