Tiered speech synthesis with over 1,000 voices, 16 languages, 20 Chinese dialects, natural-language delivery direction, and inline emotion tags.
Alibaba CloudAudio GenerationReleased Jul 20, 2026SingaporeProprietary EndpointNewTiered speech synthesis with over 1,000 voices, 16 languages, 20 Chinese dialects, natural-language delivery direction, and inline emotion tags.
Voice selection is tier aware: base voice ids work on both tiers and are automatically matched to the selected model_tier, while the six system voices are locked to one tier. Pitch also shifts the pace of the rendered audio, so pair it with speed when you want to preserve the original duration.
Also known as Qwen Audio TTS, Alibaba Cloud Qwen Audio 3.0 TTS, Qwen-Audio-3.0-TTS, qwen-audio-3-0-tts
qwen-audio-3-0-tts/v1/audio/speechPOST/v1/audio/speech:streamGET/v1/voicesqwen-audio-3.0-ttsalibaba/qwen-audio-3-0-ttsqwen-audio-3.0-tts-plusqwen-audio-3.0-tts-flashLive pay-as-you-go rates from the EmpirioLabs catalog. You are billed only for what you use, with no monthly minimum.
Qwen Audio 3.0 TTS serves speech through POST /v1/audio/speech and returns playable audio. Send the text to speak as input with the model id qwen-audio-3-0-tts. Get an API key from the EmpirioLabs dashboard.
curl https://api.empiriolabs.ai/v1/audio/speech \
-H "Authorization: Bearer $EMPIRIOLABS_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "qwen-audio-3-0-tts",
"input": "Welcome to EmpirioLabs. Your build just finished."
}' \
--output speech.mp3import requests
response = requests.post(
"https://api.empiriolabs.ai/v1/audio/speech",
headers={"Authorization": "Bearer YOUR_EMPIRIOLABS_API_KEY"},
json={"model": "qwen-audio-3-0-tts", "input": "Welcome to EmpirioLabs."},
)
with open("speech.mp3", "wb") as f:
f.write(response.content)Request parameters supported by the Qwen Audio 3.0 TTS API on EmpirioLabs. Defaults apply when a field is omitted.
| Parameter | Type | Default | Range / values | Description |
|---|---|---|---|---|
| input | string | - | max 20000 | Text to synthesize. Up to 20,000 characters per request. Supports inline expression tags placed directly in the text, for example [excited], [laughing], [whispers],... |
| model_tier | enum | plus | plus, flash | plus: highest audio quality and expressiveness, best for content creation, audiobooks, dubbing, and brand voice work. flash: tuned for real-time interaction with... |
| voice | enum | loongjameszhao | loongjameszhao, loongolivialin, loonglunawang, loongnorahu, l... | Voice preset. The base voices work on BOTH tiers, so changing model_tier keeps your selection. Six flagship voices are tier locked: longanlingxin and longanlufeng on... |
| voice_id | string | - | - | Free-form voice id, which overrides voice when set. Accepts any base voice from GET /v1/voices, either as the bare id (recommended, works on both tiers) or fully... |
| response_format | enum | mp3 | mp3, wav, pcm, opus | Audio container. mp3 and opus are compressed, wav is uncompressed PCM in a RIFF header, and pcm is headerless raw samples for chunked playback. |
| sample_rate | enum | 24000 | 8000, 16000, 22050, 24000, 44100, 48000 | Output sample rate in Hz. 24000 suits speech playback, and 48000 gives broadcast-quality output at a larger file size. |
| speed | number | 1 | 0.5 to 2 | Speaking rate multiplier. 0.5 is half speed and 2.0 is double speed. |
| pitch | number | 1 | 0.5 to 2 | Pitch multiplier. Values below 1.0 lower the voice and values above raise it. Pitch also shifts the pace of the rendered audio, so pair it with speed when you want... |
| volume | integer | 50 | 0 to 100 | Output loudness, where 50 is the reference level. |
| instruction | string | - | max 128 | Natural-language direction for delivery, up to 128 characters. Controls emotion, tone, character, pace, and speaking style, for example 'Speak quickly in an excited,... |
| language_hints | array | - | zh, en | Language codes that bias pronunciation for mixed-language text, for example ["zh", "en"]. Leave unset to let the model detect the language. |
| seed | integer | 0 | 0 to 65535 | Sampling seed. Reuse a seed with identical input and settings for a repeatable render. |
| bit_rate | integer | 32000 | 16000 to 64000 | Encoder bitrate in bps. Applies to the opus format only, and is ignored for mp3, wav, and pcm. |
| pronunciation | object | - | - | Pronunciation overrides keyed by the written form, for example {"重要": "zhong4 yao4"}. Use it to fix names, acronyms, and homographs. |
On EmpirioLabs, Qwen Audio 3.0 TTS is billed pay as you go: Plus synthesis $0.20 per 10,000 characters; Flash synthesis $0.15 per 10,000 characters. The live rate card on this page always matches what the API charges.
Qwen Audio 3.0 TTS is served through POST /v1/audio/speech on api.empiriolabs.ai with standard bearer-token authentication.
Yes. The EmpirioLabs playground runs Qwen Audio 3.0 TTS in the browser with the same parameters the API exposes, so you can test prompts before writing code.
Create an EmpirioLabs account, then generate a key under API Keys in the dashboard. Billing is pay-as-you-go credits, so you only pay for the requests you make.
Check out our pricing or reach out if you want your own model deployed on our stack.
