Library-first Chatterbox voice synthesis: text + config + seed to WAV. Ships the
tts-chatterbox Python distribution (import packages tts_chatterbox and
artifact_store) plus the CPU jobs-API service and the RunPod GPU worker.
Voice references are biometric data: cloning a voice requires the speaker's explicit
permission, and voice-reference files must never be committed to this repository or baked
into its images. This library is not a consent-free voice-cloning tool. See
docs/CONSUMPTION.md for the full consumption contract.