Qwen3-TTS
Clone a voice from a short clip, on your own CPU
Qwen3-TTS-12Hz-0.6B-Base compiled to WebAssembly with fusor-ai. Give it a few seconds of a voice, type a text in one of ten languages, and it speaks it in that voice as it computes. Once the model is loaded, neither the clip nor the text leaves the page.
Voice
or record yourself, here, in this page
✓