Qwen3-TTS

About Qwen3-TTS

Qwen3-TTS from Alibaba is a 0.6B parameter text-to-speech model that combines high quality with efficient inference. It supports 10 languages and can clone any voice from just 3 seconds of reference audio. Built on the Qwen3 architecture, it produces natural-sounding speech with excellent prosody and pronunciation across all supported languages.

Key Features

Use Cases

How to Use Qwen3-TTS

1. 1

   Sign up free or use the demo

   Create a free TextToSpeechAI account to get starter credits, or try the no-signup demo first. No GPU or local installation of Qwen3-TTS is needed - everything runs on our servers.

2. 2

   Select Qwen3-TTS and add a 3-second clip

   Choose Qwen3-TTS as your engine from the voice picker. To clone a voice, upload a clean reference clip of about 3 seconds; for a non-cloned voice, just pick one of the built-in Qwen3-TTS voices.

3. 3

   Enter your text in any of 10 languages

   Type or paste your script in Chinese, English, Japanese, Korean, French, German, Spanish, Italian, Portuguese, or Russian. Qwen3-TTS can speak your cloned voice across all 10 supported languages.

4. 4

   Generate the speech

   Click generate and Qwen3-TTS synthesizes your audio on our GPUs at the premium tier (25 credits per 1000 characters). The compact 0.6B model returns natural multilingual speech quickly.

5. 5

   Download or use the API

   Preview the result, then download the audio file or fetch it programmatically through the TextToSpeechAI API at api.texttospeechai.com. Reuse the same cloned Qwen3-TTS voice for future generations.

Qwen3-TTS API

Generate speech programmatically using the TextToSpeechAI REST API.

curl -X POST "https://api.texttospeechai.com/v1/generate/" \
  -H "Authorization: Bearer YOUR_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "text": "Qwen3\u002DTTS delivers natural multilingual speech with ultra\u002Dfast 3\u002Dsecond voice cloning.",
    "voice": "en_US-lessac-medium"
  }'

Frequently Asked Questions