Qwen-Audio-Realtime Voice configuration
After creating a cloned voice, use it in a Qwen-Audio-Realtime session and consult the corresponding HTTP API.
Use the voice cloning API to create, query, update, and delete Qwen-Audio-Realtime voices. The target_model specified during creation must match the model using the voice. Prepare samples according to Audio Requirements; see Voice Cloning User Guide for a complete example.
Creating a voice is free of charge; real-time conversations using cloned voices are billed according to the target model—see Billing Details. Voice quantity limits are outlined in Quotas and Limits.
Using Voices in Real-Time Conversations
Connect to the same Qwen-Audio-Realtime model used when creating the voice, and in the first session.update, set voice to the returned voice_id. The voice can be used once its status becomes OK.
{
"type": "session.update",
"session": {
"voice": "VOICE_ID"
}
}
Service endpoint
See Voice Cloning HTTP API: Endpoint.
Request headers
See Voice Cloning HTTP API: Request Headers.
Creating a Voice
See Voice Cloning HTTP API: Create Voice.
See Voice Cloning HTTP API: Request Body.
See Voice Cloning HTTP API: Response Body.
Query voice list
See Voice Cloning HTTP API: List Voices.
See Voice Cloning HTTP API: Request Body.
See Voice Cloning HTTP API: Response Body.
Query voice details
See Voice Cloning HTTP API: Query Voice Details.
See Voice Cloning HTTP API: Request Body.
See Voice Cloning HTTP API: Response Body.
Update a voice
See Voice Cloning HTTP API: Update Voice.
See Voice Cloning HTTP API: Request Body.
See Voice Cloning HTTP API: Response Body.
Delete a voice
See Voice Cloning HTTP API: Delete Voice.
See Voice Cloning HTTP API: Request Body.
See Voice Cloning HTTP API: Response Body.