Nodes/ComfyUI-GGUF-Loader/MiniMax Music 3 Audio Generate ⚡
ComfyUI Node

MiniMax Music 3 Audio Generate ⚡

Generate caption- and lyrics-conditioned stereo music with the native MiniMax Music 3 AR, flow DiT, and DAV pipeline.

By ChrisColeTech·Created 13 days ago·Updated about 3 hours ago· 6
MiniMax Music 3 Audio Generate ⚡
  • model
  • clip
  • vae
  • audio
captionGlobal Metadata: Dreamy synth-pop, 112 BPM, warm analog production. Vocal Details: Expressive lead vocal with layered harmonies. Arrangement: Atmospheric intro, intimate verses, wide anthemic choruses.
lyrics[intro] [verse] Write your first verse here [chorus] Write your chorus here [outro]
max_duration120.00
seed0
steps30
guidance_scale1.7
decode_modeauto
ar_cfg_scale1.5
ar_top_k50
tile_size1536
tile_overlap64
Category🤖 CCTech/MiniMax Music

Inputs (14)

NameTypeDefaultDescription
modelMODEL
clipCLIP
vaeVAE
captionSTRINGGlobal Metadata: Dreamy synth-pop, 112 BPM, warm analog production. Vocal Details: Expressive lead vocal with layered harmonies. Arrangement: Atmospheric intro, intimate verses, wide anthemic choruses.Describe genre, BPM, key, mood, production, vocal character, instruments, and section arrangement.
lyricsSTRING[intro] [verse] Write your first verse here [chorus] Write your chorus here [outro]Use tags such as [intro], [verse], [pre-chorus], [chorus], [bridge], [instrumental], [drop], and [outro]. Leave empty for an instrumental track.
max_durationFLOAT120.000.04–360Upper bound in seconds; the model may finish earlier.
seedINT00–18446744073709550000
stepsINT301–100
guidance_scaleFLOAT1.70–100
decode_modeCOMBOauto3 options: auto, tiled, dense
ar_cfg_scaleFLOAT1.50–100
ar_top_kINT501–16384
tile_sizeINT153632–8192
tile_overlapINT640–1024

Outputs (1)

NameTypeDescription
audioAUDIO