Stable Audio 3

Generate high-quality audio with Stability AI's stable-audio-3. Describe the audio you want in natural language (text-to-audio), or supply a reference clip with audioUrl to transform an existing sample into a new composition (audio-to-audio) controlled by strength.

Supported Models

• stable-audio-3 — Stable Audio 3

Parameters

Parameter	Type	Required	Description
type	string	Yes	Must be "AUDIO_GENERATION"
model	string	Yes	Must be "stable-audio-3"
promptObject	object	Yes	Stable Audio generation parameters

Prompt Object Parameters

Parameter	Type	Required	Description
prompt	string	Yes	Text description of the audio to generate
duration	number	Conditional	Length of the generated audio in seconds. min 1, max 380. Required when audioUrl is not provided
output_format	string (enum)	Yes	Output format. One of mp3, wav. Default mp3
audioUrl	string	No	Reference audio (S3 asset key — upload via the asset API first). Providing it switches to audio-to-audio mode. Input audio must be mp3/wav, 6–380s, ≤ 50 MB
strength	number	No	Audio-to-audio only. How much the input audio is transformed; higher values follow the prompt more and the input less. min 0, max 1. Default 1
steps	number	No	Number of sampling steps. Higher = more detail but slower. min 4, max 8. Default 8
cfg_scale	number	No	How strictly the audio adheres to the prompt. Higher values follow the prompt more closely. min 1, max 25. Default 1
seed	number	No	Seed for reproducible results. 0 uses a random seed. min 0, max 4294967294

Interactive API Testing

API Playground

GETPOSTPUTDELETEhttps://api.1min.ai/api/features

Prompt *lofi hip hop beat with warm vinyl crackle and a mellow piano melody

Duration (seconds)

1–380 seconds. Required when no input audio is provided

Output Format *MP3 - Compressed audio formatWAV - Uncompressed audio format

Input Audio (optional, enables audio-to-audio)

Steps

4–8. Higher = more detail but slower

Guidance Scale

1–25. Higher follows the prompt more closely

Seed (optional)

0 uses a random seed

Generated cURL Command:

curl -X POST "https://api.1min.ai/api/features" \
  -H "API-KEY: <your-api-key>" \
  -H "Content-Type: application/json" \
  -d '{
  "type": "AUDIO_GENERATION",
  "model": "stable-audio-3",
  "conversationId": "AUDIO_GENERATION",
  "promptObject": {
    "prompt": "lofi hip hop beat with warm vinyl crackle and a mellow piano melody",
    "duration": 60,
    "output_format": "mp3",
    "steps": 8,
    "cfg_scale": 1
  }
}'

Response Format

Successful requests return an aiRecord echoing the submitted promptObject and a resultObject containing the generated audio file path.

{
  "aiRecord": {
    "uuid": "9f4a2b1e-6c3d-4f5e-9a8b-2d1c0e3f4a5b",
    "userId": "c937fbcc-fa8f-4565-a440-c4d87f56fcb2",
    "teamId": "a4e176b2-dabb-451e-9c58-62b451fa9630",
    "model": "stable-audio-3",
    "type": "AUDIO_GENERATION",
    "status": "SUCCESS",
    "createdAt": "2026-06-20T10:52:19.827Z",
    "aiRecordDetail": {
      "promptObject": {
        "prompt": "lofi hip hop beat with warm vinyl crackle and a mellow piano melody",
        "duration": 60,
        "output_format": "mp3",
        "steps": 8,
        "cfg_scale": 1
      },
      "responseObject": {
        "finish_reason": "SUCCESS",
        "seed": 12345
      },
      "resultObject": [
        "development/audios/2026_06_20_17_52_25_414_538984.mp3"
      ]
    },
    "temporaryUrl": ""
  }
}

Example Requests

Text-to-audio

{
  "type": "AUDIO_GENERATION",
  "model": "stable-audio-3",
  "promptObject": {
    "prompt": "energetic electronic dance track with pulsing synthesizers, heavy bass drops, and crisp hi-hats",
    "duration": 30,
    "output_format": "wav",
    "steps": 8,
    "cfg_scale": 5,
    "seed": 42
  }
}

Audio-to-audio

Upload the reference audio to the asset API first; pass the returned S3 key as audioUrl.

{
  "type": "AUDIO_GENERATION",
  "model": "stable-audio-3",
  "promptObject": {
    "prompt": "transform this loop into a cinematic orchestral build with strings and brass",
    "audioUrl": "development/audios/2026_06_20_12_31_05_a1b2c3.mp3",
    "strength": 0.65,
    "output_format": "mp3",
    "steps": 8,
    "cfg_scale": 3
  }
}