---
library_name: mlx-audio
base_model:
- ResembleAI/chatterbox
language:
- en
- ar
- hi
- de
- es
- fr
- it
- pt
- ru
- ja
- ko
- zh
license: mit
tags:
- mlx
- text-to-speech
- voice-cloning
- multilingual
pipeline_tag: text-to-speech
---

# aufklarer/Chatterbox-Multilingual-MLX-fp16

Multilingual [Chatterbox](https://huggingface.co/ResembleAI/chatterbox) (Resemble AI,
MIT) converted to MLX in **genuine fp16** for Apple-silicon inference. Unlike
`mlx-community/chatterbox-fp16` (stored as F32, ~2.6 GB), this bundle stores all
float tensors as fp16 (~1.3 GB) with an identical key layout.

Zero-shot voice cloning from a short reference clip across 23 languages
(incl. Arabic and Hindi). Component prefixes: `ve.*` (voice encoder), `t3.*`
(text→speech-token T3), `s3gen.*` (token→waveform S3Gen).

**Note:** requires the S3Tokenizer weights from
[mlx-community/S3TokenizerV2](https://huggingface.co/mlx-community/S3TokenizerV2),
downloaded automatically at runtime.

## Use with mlx-audio

```bash
pip install -U mlx-audio
mlx_audio.tts.generate --model aufklarer/Chatterbox-Multilingual-MLX-fp16 --text "[ar] مرحبا" --ref_audio reference.wav
```

Converted with `models/chatterbox/export/convert.py` (speech-models).