vox-chatterbox

vox-chatterbox provides Vox TTS adapters for Resemble AI Chatterbox.

Adapters:

• chatterbox-tts-turbo - Chatterbox Turbo backend
• chatterbox-tts-turbo-onnx - Chatterbox Turbo ONNX backend
• chatterbox-tts - Chatterbox backend
• chatterbox-tts-multilingual - Chatterbox multilingual backend

Install

pip install vox-chatterbox

Runtime Dependencies

The adapter package is intentionally light. The Chatterbox backend package is installed on demand into the isolated target runtime $VOX_HOME/runtime/chatterbox.

The ONNX Turbo backend uses the same runtime directory and installs ONNX Runtime plus tokenizer/audio helpers on demand. The official Chatterbox Turbo ONNX graph requires speaker conditioning, so this backend requires reference_audio or a voice value pointing to a WAV file.

Use with Vox

vox pull chatterbox-tts-turbo:0.1.7
vox run chatterbox-tts-turbo:0.1.7 "Hello from Chatterbox"
vox pull chatterbox-tts-turbo:0.1.7 --variant onnx

For voice cloning, pass a reference audio sample through the Vox API or use a voice value that points to a local WAV file.

Long input is split at sentence or clause boundaries before synthesis. Each backend call receives at most 200 characters, preventing the coherence drift that Chatterbox can develop near the end of long single-shot generations. Vox returns the resulting audio as one continuous response.