AI voice models and products powering millions of developers, creators and enterprises. Generate lifelike speech from text in seconds — in studio quality, at scale.
Closed beta · access by invite only
We're onboarding a small group of creators and developers while we polish the platform. Have an invite? Sign in and start generating.
One platform for speech synthesis — from a single narration to millions of API calls.
Qwen3-TTS renders speech with human rhythm, emotion and emphasis — indistinguishable from a studio recording.
Chinese, English, Japanese, Korean and more — plus Beijing and Sichuan dialect speakers.
The open-source 1.7B model synthesizes on-device — free, private, and with no per-character fees.
9 curated voices across narration, ads, news and entertainment — each with its own character.
A clean REST API with per-key usage metering. Ship voice into your product with a five-line curl call.
API key management, quota enforcement, full generation history and usage analytics out of the box.
The 9 official Qwen3-TTS CustomVoice speakers — narrators, dialect personalities and international characters, all running on-device.
Warm, gentle young female voice. Calm and soothing — ideal for narration and audiobooks.
Bright, slightly edgy young female voice. Confident, cute, and a little feisty.
Seasoned male voice with a low, mellow timbre. Unhurried gravitas for storytelling.
Dynamic male voice with strong rhythmic drive. Dramatic flair, trailer-grade energy.
Sunny American male voice with a clear midrange. Easy-going and conversational.
Youthful Beijing male voice with a clear, natural timbre — raised in the hutongs.
Lively Chengdu male voice with a slightly husky brightness.
Playful Japanese female voice with a light, nimble timbre — a clever childhood friend.
A developer-first REST API backed by Qwen-TTS. Create an API key in the dashboard, pick a voice, and stream studio-quality audio into your app.
curl -X POST \
https://your-host/api/v1/text-to-speech/serena \
-H "xi-api-key: qk_..." \
-H "Content-Type: application/json" \
-d '{
"text": "Hello from Audionana!",
"model_id": "qwen3-tts-local",
"voice_settings": { "speed": 1.0, "pitch": 0 }
}' \
--output speech.mp3Audionana is in closed beta — sign-in is by invite only. Got a code? The studio is one click away.
Sign in to the Studio