Sign in
Recipes

Generate a complete song from a structured caption and lyrics with MiniMax Music 3.

Changed on
post/v2/consumer/recipes/miniMaxMusic3

Request

  • The document declares no server URL.
  • Auth: none required

Query parameters

experimentalboolean
allowMatureContentboolean
whatifboolean
ephemeralboolean
tier'paid' | 'free'

How a workflow is paid for.

Request body

captionstring required

Structured music description. For best results use Global Metadata, Vocal Details, and Arrangement sections.

lyricsstring required

Lyrics with section markers such as [Intro], [Verse], [Chorus], and [Outro].

seedinteger required
maxDurationnumber float

Maximum generated song duration in seconds. The model may end the song earlier when the requested lyric structure is complete.

stepsinteger
cfgnumber float

Classifier-free guidance scale used by both text encoding and sampling.

topKinteger

Top-k sampling limit used by the MiniMax music text encoder.

diffusionModelstring nullable
textEncoderstring nullable
vaestring nullable
lorasobject

Optional LoRAs to apply to both the diffusion model and MiniMax text encoder. Compatibility with the selected base resources is the caller's responsibility.

Response

OK

Changes