ACE-Step

https://mcp.app.layer.ai/mcp

execute_forge({
  "base_model_id": "ace-step",
  "prompt": "A heavy stone door grinding open in a dungeon",
  "batch_size": 1
})

# Start the generation — returns an inference id straight away.
curl -X POST https://api.app.layer.ai/api/v2/workspaces/$WORKSPACE_ID/base-models/ace-step/inferences \
  -H "Authorization: Bearer $LAYER_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
        "prompt": "A heavy stone door grinding open in a dungeon"
      }'

# Poll until status is "complete" — the response then carries the asset URLs.
curl https://api.app.layer.ai/api/v2/workspaces/$WORKSPACE_ID/inferences/$INFERENCE_ID \
  -H "Authorization: Bearer $LAYER_TOKEN"

prompt

string

Describe what to generate, or the change to make.

seed

integer

The starting point for the model's randomness. The same seed and settings give the same result.

guidance_scale

numberdefault 150–200

How closely the result follows the prompt. Lower gives the model more freedom.

duration_seconds

numberdefault 605–240

Length of the generated clip, in seconds.

guidance_file_init_audio

file[]

The audio to modify

guidance_file_reference_audio

file[]

Guides style and voice without being modified

extend_seconds

number1–60

How much new audio continues the clip past its end.

audio_sections

array

The stretches of the clip the prompt rewrites; the rest is kept.

instrumental

boolean

Music only, with no vocals.

prompt_language

string

The language the prompt is written in, so it reaches the model as you meant it.