---
license: apache-2.0
tags:
  - thestage
  - qwen3-tts
  - voice-clone
  - apple
  - ios
language:
  - en
  - es
  - fr
  - de
  - pt
---

# Qwen3-TTS Tutor Voice Packs

Small **per-language clone packs** for
[`TheStageAI/Qwen3-TTS-12Hz-0.6B-Base`](https://huggingface.co/TheStageAI/Qwen3-TTS-12Hz-0.6B-Base)
on Apple Silicon (TheStage Apple SDK).

Used by the **TutorTTS** mixed-language demo (`examples/tutor_tts` in
[AppleSDK](https://github.com/TheStageAI/AppleSDK)): tagged scripts like
`<en>…</en><es>…</es>` call `set_voice` between spans.

## Layout

```
tutor_en/voice.json
tutor_es/voice.json
tutor_fr/voice.json
tutor_de/voice.json
tutor_pt/voice.json
```

Each `voice.json` is a Qwen3-TTS Base clone pack (ref codes + metadata).

## Languages

| Folder | Tag in scripts |
|--------|----------------|
| `tutor_en` | `<en>` |
| `tutor_es` | `<es>` |
| `tutor_fr` | `<fr>` |
| `tutor_de` | `<de>` |
| `tutor_pt` | `<pt>` |

## Download (demo)

Packs are tiny (~30 KB each `voice.json`) — HF is for easy hub cache /
`hf download`, not because the blobs are huge:

```bash
hf download TheStageAI/Qwen3-TTS-Tutor-VoicePacks --local-dir ./VoicePacks
```

Point TutorTTS `voice_dir` at the downloaded folder (or copy `tutor_*`
into the app's `VoicePacks/`).

## License

Apache-2.0. Packs are derived for demo use with TheStage Qwen3-TTS Base.