README.md

---
title: 周面範 Voice Clone
emoji: 🎙️
colorFrom: red
colorTo: yellow
sdk: gradio
sdk_version: 4.44.1
app_file: app.py
pinned: false
python_version: "3.11"
short_description: Multilingual voice cloning with Cantonese support
---

# 周面範 · Voice Clone (Hugging Face Space)

This Space hosts the Gradio UI for the voice-cloning project. Synthesis
runs on two backends:

- **Chatterbox Multilingual** on Modal for en / zh / ja / ko / fr / es.
- **MiniMax Speech-2.8** for 粵語 Cantonese (mandatory) via
  `language_boost="Chinese,Yue"`.

## Required secrets (Settings → Variables and secrets)

| Secret | Required | Purpose |
|---|---|---|
| `MINIMAX_API_KEY` | yes | Cantonese synthesis |
| `CHATTERBOX_API_BASE_URL` | yes | Your Modal endpoint URL |
| `MINIMAX_GROUP_ID` | optional | Only if your MiniMax region requires it |
| `CHATTERBOX_API_KEY` | optional | Only if your Modal deploy adds auth |

After setting the secrets, the Space will automatically restart and
serve the Gradio UI at `https://<owner>-<space-name>.hf.space`.

## Source

Full source (tests, Modal deploy, docs): accompanying repository.