Speech understanding, generation and interaction research from Alibaba. The Qwen Audio homepage introduces the Qwen-Audio series and connects each project to its demonstrations and research materials.
| Project | Focus | Project page |
|---|---|---|
| Qwen-Audio-3.0-ASR | Multilingual and dialect recognition, streaming, contextual recognition and hotwords | Explore ASR |
| Qwen-Audio-3.0-TTS | Controllable speech synthesis, cross-lingual voice cloning and long-form generation | Explore TTS |
| Project | Project page |
|---|---|
| SenseVoice & CosyVoice (original FunAudioLLM project) | https://qwenaudio.github.io/sensevoice-cosyvoice/ |
| CosyVoice 2 | https://qwenaudio.github.io/cosyvoice2/ |
| CosyVoice 3 | https://qwenaudio.github.io/cosyvoice3/ |
| MinMo | https://qwenaudio.github.io/minmo/ |
| InspireMusic | https://qwenaudio.github.io/inspiremusic/ |
| Fun-ASR | https://qwenaudio.github.io/funasr/ |
| Fun-Audio-Chat | https://qwenaudio.github.io/funaudiochat/ |
index.htmlandhub/: bilingual Qwen Audio project directory.sensevoice-cosyvoice/index.html: the original SenseVoice & CosyVoice page.- Other project directories retain their existing pages and URLs.
- Shared original-page assets remain in
audios/,pics/,css/andpdf/. The moved page uses the site root as its asset base; its contents links point directly to its new URL. Previously shared root fragment links are forwarded to the corresponding original project section.
The site is static and published with GitHub Pages. No build step is required.
For local preview, serve this directory as the web root, for example with
python -m http.server 8000.