ποΈ Qwen3-TTS: 1-click launcher with voice cloning
Run Alibaba Qwen's new open-source text-to-speech models locally with one click.
𧬠Clone a voice with a short clip
Upload 3+ seconds of anyone's voice, type what they said, and it speaks any text in that voice. You can save a cloned
voice as a file and reuse it later, with no need to upload the clip again.
π£οΈ 9 built-in voices
Vivian, Serena, Uncle Fu, Dylan, Eric, Ryan, Aiden, Ono Anna and Sohee. Add a style instruction like "Say it in a very
angry tone" or "Very happy."
π¨ Voice Design
Describe a voice in plain words, like "a calm, elderly male narrator with a deep voice", and the model creates it.
π 10 languages
Chinese, English, Japanese, Korean, German, French, Russian, Portuguese, Spanish and Italian, plus Beijing and Sichuan
dialect voices.
ποΈ Studio: everything in one UI
Voice Clone, Custom Voice and Voice Design are tabs in one app. Switch between 1.7B (best quality) and 0.6B (faster,
lighter). Models download the first time you use them, and only one is loaded at a time to save VRAM.
π§° Also included
- Each model can still be started on its own in the official Qwen demo.
- Every button is also an API (Python, JavaScript, curl), so you can call it from your own scripts.
- Runs only on your own computer (127.0.0.1).
- Works on NVIDIA, AMD (Linux), Apple Silicon and CPU. 8 GB+ VRAM is recommended.
βΆοΈ How to use
- Install
- Start Studio
- Open the Voice Clone tab, drop in a clip and type its transcript
- Type your text and click Generate
π https://github.com/PierrunoYT/Qwen3-TTS-Pinokio
Models: Apache 2.0 (Qwen). Launcher: MIT.
β οΈ Only clone voices you have permission to use.
ποΈ Qwen3-TTS: 1-click launcher with voice cloning
Run Alibaba Qwen's new open-source text-to-speech models locally with one click.
𧬠Clone a voice with a short clip
Upload 3+ seconds of anyone's voice, type what they said, and it speaks any text in that voice. You can save a cloned
voice as a file and reuse it later, with no need to upload the clip again.
π£οΈ 9 built-in voices
Vivian, Serena, Uncle Fu, Dylan, Eric, Ryan, Aiden, Ono Anna and Sohee. Add a style instruction like "Say it in a very
angry tone" or "Very happy."
π¨ Voice Design
Describe a voice in plain words, like "a calm, elderly male narrator with a deep voice", and the model creates it.
π 10 languages
Chinese, English, Japanese, Korean, German, French, Russian, Portuguese, Spanish and Italian, plus Beijing and Sichuan
dialect voices.
ποΈ Studio: everything in one UI
Voice Clone, Custom Voice and Voice Design are tabs in one app. Switch between 1.7B (best quality) and 0.6B (faster,
lighter). Models download the first time you use them, and only one is loaded at a time to save VRAM.
π§° Also included
βΆοΈ How to use
π https://github.com/PierrunoYT/Qwen3-TTS-Pinokio
Models: Apache 2.0 (Qwen). Launcher: MIT.
β οΈ Only clone voices you have permission to use.