I built a way to test Qwen3-TTS and Qwen3-ASR locally on your laptop

Supports Qwen3-TTS models (0.6B-1.7B) and ASR models. Docker + native deployment options.

Key features:

• 🎭 Voice cloning with reference audio
• 🎨 Custom voice design from text descriptions
• ⚡ MLX + Metal GPU acceleration for M1/M2/M3
• 🎨 Modern React UI included

If you like local audio models, give it a try. Works best in local dev mode for now.