Music and cover generation

This API generates music or a cover within a single synchronous HTTP request. A standard request can return a temporary audio URL or Hex-encoded audio content; a streaming request returns JSON audio chunks line by line.

For a streaming result, concatenate the Hex content in data.audio in the order received, then decode and save it after receiving data.status=2. Even when the HTTP status is 200, check base_resp.status_code; only 0 indicates business-level success.

POST/v1/music_generation

Generate music or a cover synchronously. When stream=true, the response returns JSON audio chunks line by line over the same HTTP connection and does not create an asynchronous task.

In a streaming result, data.audio contains Hex audio chunks that must be concatenated in the order received. data.status=2 indicates that generation is complete.

Streaming responses always return Hex audio chunks. When stream=true, the current gateway may ignore output_format=url and still return Hex chunks.

Authorization

bearerAuth

AuthorizationBearer <token>

Use Bearer your-api-key in the Authorization Header.

In: header

Request Body

application/json

Music or cover generation parameters. Conditional requirements depend on the generation mode.

TypeScript Definitions

Use the request body type in TypeScript.

model*string

MiniMax music model ID visible to the current API Key, such as music-3.0, music-2.6, or music-cover.

prompt?string

Style, mood, and scene description. Required in instrumental mode and optional otherwise.

Lengthlength <= 2000

lyrics?string

Lyrics separated by line breaks. Usually required for non-instrumental music; they can be generated automatically when lyric optimization is enabled and this field is empty.

Lengthlength <= 3500

stream?boolean

Whether to return JSON Hex audio chunks line by line in the current HTTP response.

Defaultfalse

output_format?string

Audio return method. Streaming responses always use hex; the current gateway may ignore url and still return Hex chunks. URLs usually expire after 24 hours.

Default"hex"

Value in

"url"
"hex"

audio_setting?

lyrics_optimizer?boolean

Whether to generate lyrics from the prompt when the lyrics are empty.

Defaultfalse

is_instrumental?boolean

Whether to generate instrumental music without vocals. When true, lyrics is not required.

Defaultfalse

audio_url?string

Reference audio URL for a cover; mutually exclusive with audio_base64 and cover_feature_id.

Formaturi

audio_base64?string

Base64 reference audio for a cover; mutually exclusive with audio_url and cover_feature_id.

cover_feature_id?string

Feature ID returned by cover preprocessing, valid for 24 hours. When using it, provide lyrics containing 10 to 1000 characters.

aigc_watermark?boolean

Whether to add an AIGC watermark at the end of the audio. Applies only to non-streaming requests.

Defaultfalse

Response Body

`application/json`

curl -X POST "https://example.com/v1/music_generation" \  -H "Content-Type: application/json" \  -d '{    "model": "MODEL_ID",    "prompt": "Gentle piano and strings, sunrise at dawn, warm and uplifting, cinematic score style",    "is_instrumental": true,    "audio_setting": {      "sample_rate": 44100,      "bitrate": 256000,      "format": "mp3"    },    "output_format": "url",    "stream": false  }'

{
  "data": {
    "audio": "https://example.com/music/output.mp3",
    "status": 2
  },
  "trace_id": "04ede0ab069fb1ba8be5156a24b1e081",
  "extra_info": {
    "music_duration": 244262,
    "music_sample_rate": 44100,
    "music_channel": 2,
    "bitrate": 256000,
    "music_size": 7818388
  },
  "analysis_info": null,
  "base_resp": {
    "status_code": 0,
    "status_msg": "success"
  }
}

Authorization

Request Body

Response Body

200application/json

401application/json

`application/json`

`application/json`