validateInfo text returned by the server. For best voice generation results, have the user record the server-provided phrase in a singing voice rather than plain speech.taskId and the verification audio URL in verifyUrl.voiceName, description, style, and singerSkillLevel helps organize and tune the generated voice.taskId; use it to query the final voiceId.callBackUrl is provided, the system sends a POST callback when the voice is created or the task fails. The callback URL must be publicly accessible and return HTTP 200 within 15 seconds.verifyUrl.taskId.voiceId through the record query API or callback when the task succeeds.taskId must come from the validation phrase task for the same voice workflow.verifyUrl should point to the user's recording of the exact validation phrase returned by the server; for best results, recording it in a singing voice is recommended.voiceId, use the availability check endpoint before starting generation workflows that depend on the custom voice.curl --location 'https://apibox.erweima.ai/api/v1/voice/generate' \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '{
"taskId": "voice_create_001",
"verifyUrl": "https://example.com/audio/verify_read.mp3",
"voiceName": "My Voice",
"description": "created from uploaded voice",
"style": "Pop, Female Vocal",
"singerSkillLevel": "beginner",
"callBackUrl": "https://example.com/callback/suno/voice_create"
}'{
"code": 200,
"msg": "success",
"data": {
"taskId": "xxx_task_id_xxx"
}
}