Qwen3 Tts Voice Clone

Qwen3 Tts Voice Clone

Playground

Qwen3 TTS Voice Clone: Clone any voice from a reference audio and generate speech in that voice. Ready-to-use REST inference API, best performance, no cold starts, affordable pricing.

Features

Qwen3-TTS Voice Clone

Qwen3-TTS Voice Clone is an advanced text-to-speech model that clones voices from reference audio. Upload a short audio sample of any voice, and the model generates new speech in that exact voice — preserving tone, accent, and speaking style.

Why Choose This?

• High-fidelity voice cloning Capture the unique characteristics of any voice from just a short audio sample.

• Reference transcript support Provide the transcript of your reference audio to improve cloning accuracy.

• Multilingual support Generate cloned voice speech in 10 languages: Chinese, English, German, Italian, Portuguese, Spanish, Japanese, Korean, French, and Russian.

• Auto language detection Set language to “auto” and the model intelligently detects the language from your text.

Parameters

How to Use

1. Upload reference audio — provide a clear audio sample of the voice you want to clone (3-15 seconds recommended).
2. Add reference transcript (optional) — enter the exact text spoken in your reference audio to improve cloning accuracy.
3. Enter your text — write or paste the content you want to convert to speech.
4. Select language — choose the target language or use “auto” for automatic detection.
5. Run — submit and download your audio file.

Pricing

Billing Rules

• Minimum charge: $0.02 (for texts under 1,000 characters)
• For longer texts: $0.02 × (character count / 1,000)

Best Use Cases

• Personalized Voiceovers — Clone your own voice to generate content without recording.
• Character Consistency — Maintain the same voice across multiple audio productions.
• Localization — Clone a voice to speak in different languages while preserving identity.
• Audiobook Production — Generate hours of narration from a single voice sample.
• Accessibility — Create personalized text-to-speech voices for individuals.

Pro Tips

• Use clean, noise-free reference audio for best cloning results.
• Reference audio of 3-15 seconds with clear speech works best.
• Always provide reference_text when possible — it significantly improves voice matching accuracy.
• Ensure the reference audio contains natural speech without music or background noise.
• The cloned voice works best when the target text matches the reference audio’s language.

Related Models

• Qwen3-TTS Voice Design — Design custom voices using natural language descriptions instead of audio samples.

Notes

• Reference audio quality directly affects cloning quality — use high-quality recordings.
• The model preserves accent, tone, and speaking style from the reference.
• For best results, match the language parameter to your text content.

# Submit the task
curl --location --request POST "https://api.wavespeed.ai/api/v3/wavespeed-ai/qwen3-tts/voice-clone" \
--header "Content-Type: application/json" \
--header "Authorization: Bearer ${WAVESPEED_API_KEY}" \
--data-raw '{
    "language": "auto"
}'

# Get the result
curl --location --request GET "https://api.wavespeed.ai/api/v3/predictions/${requestId}/result" \
--header "Authorization: Bearer ${WAVESPEED_API_KEY}"