Skip to content

Mureka โ€‹

Updated

Endpoint: POST https://vod.intl.tencentcloudapi.comAction: CreateAigcAudioTask

Mureka is a music generation model with three modes: lyrics to song, prompt to song, and instrumental generation. Versions song_8, song_9, song_9.5, easy_song_9.5, and instrumental_9.5 are available; the three 9.5 versions were launched on 2026-11-16.

See Version differences to pick a version and Calling by version for request examples.

Basic Information โ€‹

ItemValue
ModelNameMureka
ModelVersionsong_8 / song_9 / song_9.5 / easy_song_9.5 / instrumental_9.5
SceneTypeNot required โ€” leave it empty
UseLyrics to song (song_*) / prompt to song (easy_song_9.5) / instrumental (instrumental_9.5)

Version differences โ€‹

ModelVersionModeLyrics (lyrics)VocalsUpstream Mureka API
song_8Lyrics to songโœ…โœ…Lyrics to song
song_9Lyrics to songโœ…โœ…Same as above
song_9.5Lyrics to songโœ…โœ…Same as above
easy_song_9.5Prompt to song onlyโ€”โœ…Prompt to song
instrumental_9.5Instrumentalโ€”โ€” (no vocals)Generate instrumental

How to choose:

  • You have lyrics โ†’ song_9.5 (the latest); song_8 / song_9 remain available.
  • You only have a description and want the model to write the song โ†’ easy_song_9.5.
  • Background music with no vocals โ†’ instrumental_9.5.

The integration guide only states that easy_song_9.5 "supports prompt-to-song only" and instrumental_9.5 "generates instrumentals". The "Lyrics" and "Vocals" columns above are inferred from the corresponding upstream Mureka APIs, which have no lyrics field for those two endpoints. Do not pass lyrics to them.

Request Parameters โ€‹

ParameterRequiredTypeDescription
SubAppIdOptionalIntegerVOD application ID
ModelNameOptionalStringMureka
ModelVersionOptionalStringsong_8 / song_9 / song_9.5 / easy_song_9.5 / instrumental_9.5
SceneTypeOptionalStringNot required
PromptOptionalStringStyle / instrument / mood / scene description
OutputConfigOptionalObjectOutput settings, see below
AdditionalParametersOptionalStringJSON string. Lyrics go in lyrics (song_* only)

OutputConfig โ€‹

FieldDescription
StorageModeTemporary: not stored in VOD; a temporary URL valid for 7 days is returned in the event notification. Permanent: stored in VOD; a FileId is returned
OutputAudioFormatOutput format: mp3 / wav / flac. Leave empty to use the default

Calling by version โ€‹

Lyrics to song (song_9.5) โ€‹

Same call for song_8 / song_9 โ€” just change ModelVersion.

json
{
  "ModelName": "Mureka",
  "ModelVersion": "song_9.5",
  "Prompt": "Upbeat synth-pop, female vocal, bright and catchy chorus",
  "SubAppId": 123456789,
  "OutputConfig": {
    "StorageMode": "Temporary",
    "OutputAudioFormat": "wav"
  },
  "AdditionalParameters": "{\"lyrics\":\"City lights fade, we find our way\"}"
}

Prompt to song (easy_song_9.5) โ€‹

Describe the song in Prompt only; the model writes the lyrics. No AdditionalParameters.lyrics.

json
{
  "ModelName": "Mureka",
  "ModelVersion": "easy_song_9.5",
  "Prompt": "A bright summer love song with a catchy chorus",
  "SubAppId": 123456789,
  "OutputConfig": {
    "StorageMode": "Temporary",
    "OutputAudioFormat": "mp3"
  }
}

Instrumental (instrumental_9.5) โ€‹

json
{
  "ModelName": "Mureka",
  "ModelVersion": "instrumental_9.5",
  "Prompt": "A quiet electronic instrumental for a night drive, slow tempo, warm pads",
  "SubAppId": 123456789,
  "OutputConfig": {
    "StorageMode": "Permanent",
    "OutputAudioFormat": "flac"
  }
}

Response Example โ€‹

json
{
  "Response": {
    "TaskId": "<task-id>",
    "RequestId": "<request-id>"
  }
}

Query the completion result with DescribeTaskDetail. See Audio Result Structure.

Special Settings โ€‹

  • Lyrics are specified in the AdditionalParameters JSON string's lyrics, and only apply to song_8 / song_9 / song_9.5. Do not confuse this with MPS MiniMaxMusic's lyric field.
  • AdditionalParameters is a JSON string, so the inner quotes must be escaped ("{\"lyrics\":\"...\"}").
  • Output formats are mp3 / wav / flac.
  • Prompt length (reference values from the upstream Mureka API, not stated in the VOD guide): prompt to song up to 2000 characters, instrumental up to 1024 characters.

Built with VitePress ยท Deployed on Tencent EdgeOne Pages