Inputs:
- Text prompt
- Voice (optional)
- Language (optional)
- Style prompt (optional)
- Speaker voice embedding (optional)
- Reference text (optional)
Limits:
- Audio output only
- Supported languages: Auto, English, Chinese, Spanish, French, German, Italian, Japanese, Korean, Portuguese, Russian
- Voice cloning requires a speaker embedding from Qwen 3 TTS Clone Voice
- Default output sample rate: 24 kHz
Tips:
- Use it for narration, character speech, multilingual voiceovers, and stylized TTS
- Use the style prompt to guide tone, emotion, or delivery
- Use predefined voices for fast generation
- Use speaker voice embedding when voice consistency matters
- Keep text clean and punctuated for better pacing