Errors

When a request fails to generate audio, the API returns a JSON error response containing error_code, error_message, succeed, and data. The error_code is an application-level error code, not an HTTP status code. The language of error_message follows the Accept-Language header: English by default and Chinese when Chinese is requested. Error responses do not include a trace_id.

{
  "error_code": 10003,
  "error_message": "Unauthorized",
  "succeed": false,
  "data": null
}

HTTP Status	Business error_code	Meaning	What to Do
400	10001	Invalid request content or unsupported option	Check the request and try again
400	10013	Text exceeds the 500 weighted-character limit	Shorten the text; each Chinese character counts as 2 and each other Unicode character counts as 1
400	10023	No synthesizable content	Please provide valid text or a supported event tag
401	10003	Missing or invalid API key	Check the API key
402	20003	Insufficient account balance	Add funds and try again
403	10004	Text to Speech is unavailable for the current project	Check the project’s enabled capabilities
429	10005 or 10014	The request limit has been reached	Wait briefly and try again
500	10000	An unexpected service error occurred	Try again; contact support if the problem continues
503	20002	The service is temporarily unavailable	Try again later
504	10006	The service timed out before audio was returned	Try again later

Voice Cloning errors

Voice-management endpoints return HTTP 200 on success. Failures use a non-2xx HTTP status and the standard JSON envelope. Check the HTTP status together with succeed and error_code.

HTTP Status	Business error_code	Meaning	What to Do
409	10017	The custom-voice quota has been reached	Delete an unused custom voice and try again
409	10018	voice_id already exists or is reserved	Use another voice_id
409	10019	The generated or requested voice ID conflicts with an existing voice	Retry or use another voice_id
413	20010	File exceeds the size limit for this API	File upload: up to 50 MiB; voice cloning reference audio: up to 20 MiB
400	20012	Reference audio format is invalid or cannot be processed, including corrupt media or a missing audio stream	Upload a supported, readable reference audio file
400	20015	Reference audio exceeds 60 seconds	Upload a reference recording between 3 and 60 seconds
400	20016	Reference audio is shorter than 3 seconds or does not contain enough usable speech	Upload clear, continuous speech rather than silence or noise
400	20017	Pauses in the reference speech are too long	Upload a recording with more continuous speech

If streaming generation stops unexpectedly, provide the HTTP status, API endpoint, and approximate time with time zone when contacting support.