{
  "family": "stable_audio",
  "display_name": "Stable Audio 3",
  "description": "Stability AI generative audio model family for text-to-music, sound effects, audio-to-audio editing, inpainting, continuation, variable-length generation, and LoRA personalization.",
  "category": "audio_generation",
  "status": "supported",
  "tasks": [
    "music",
    "sfx",
    "edit"
  ],
  "modes": [
    "offline"
  ],
  "languages": [
    "en"
  ],
  "options": {
    "request": [
      {
        "name": "prompt",
        "type": "string",
        "description": "Fallback text prompt when text_input is absent; pipe separates batch prompts.",
        "required": false
      },
      {
        "name": "negative_prompt",
        "type": "string",
        "description": "Negative prompt; pipe separates batch prompts.",
        "required": false
      },
      {
        "name": "duration_sec",
        "type": "float",
        "description": "Positive duration in seconds; batch requests also accept a comma-separated string matching the prompt count, not a JSON array.",
        "required": false,
        "default": 120
      },
      {
        "name": "batch_size",
        "type": "int",
        "description": "Prompt batch size.",
        "required": false,
        "min": 1,
        "default": 1
      },
      {
        "name": "num_inference_steps",
        "type": "int",
        "description": "Diffusion steps.",
        "required": false,
        "min": 1,
        "default": 8
      },
      {
        "name": "guidance_scale",
        "type": "float",
        "description": "Classifier-free guidance scale.",
        "required": false,
        "default": 1
      },
      {
        "name": "apg_scale",
        "type": "float",
        "description": "Adaptive projected guidance scale.",
        "required": false,
        "default": 1
      },
      {
        "name": "sampler_mode",
        "type": "enum",
        "description": "Stable Audio 3 diffusion sampler.",
        "required": false,
        "values": [
          "pingpong",
          "euler"
        ]
      },
      {
        "name": "seed",
        "type": "int",
        "description": "Sampling seed; omitted chooses a random seed.",
        "required": false,
        "min": 0
      },
      {
        "name": "duration_padding_sec",
        "type": "float",
        "description": "Generated padding before truncation.",
        "required": false,
        "min": 0,
        "default": 6
      },
      {
        "name": "truncate_output_to_duration",
        "type": "bool",
        "description": "Trim decoded audio to requested duration.",
        "required": false,
        "default": true
      },
      {
        "name": "chunked_decode",
        "type": "bool",
        "description": "Decode audio in chunks.",
        "required": false,
        "default": true
      },
      {
        "name": "init_noise_level",
        "type": "float",
        "description": "Audio-conditioned generation strength in [0, 1].",
        "required": false,
        "min": 0,
        "max": 1,
        "default": 1
      },
      {
        "name": "audio_input_kind",
        "type": "enum",
        "description": "How to use audio_input.",
        "required": false,
        "values": [
          "init_audio",
          "inpaint_audio"
        ],
        "default": "init_audio"
      },
      {
        "name": "inpaint_mask_start_sec",
        "type": "string",
        "description": "Comma-separated inpaint region starts in seconds.",
        "required": false
      },
      {
        "name": "inpaint_mask_end_sec",
        "type": "string",
        "description": "Comma-separated inpaint region ends in seconds.",
        "required": false
      }
    ],
    "session": [
      {
        "name": "max_batch",
        "type": "int",
        "description": "Maximum prompt batch size.",
        "required": false,
        "min": 1,
        "default": 1
      },
      {
        "name": "weight_type",
        "type": "enum",
        "description": "Weight storage type; defaults to native.",
        "required": false,
        "preset": "weight_type_full"
      },
      {
        "name": "mem_saver",
        "type": "bool",
        "description": "Release staged runtime graphs after each request.",
        "required": false,
        "default": false
      }
    ],
    "load": []
  },
  "capabilities": {
    "music": [
      "lyrics"
    ],
    "sfx": [
      "prompt_generation"
    ],
    "edit": [
      "prompt_editing"
    ]
  },
  "runtime": {
    "tags": [
      "gguf"
    ]
  },
  "ui": {
    "recommended_package": "stable_audio_3_medium_q8_0",
    "tags": [
      "Music",
      "SFX",
      "Edit",
      "GGUF"
    ],
    "docs": [
      "docs/models/stable_audio.md",
      "docs/music_generation.md",
      "docs/gguf.md"
    ]
  },
  "package_defaults": {
    "download": {
      "kind": "huggingface_snapshot",
      "repo": "audio-cpp/audio.cpp-gguf",
      "revision": "main",
      "gated": false
    }
  },
  "packages": [
    {
      "id": "stable_audio_3_medium_q8_0",
      "display_name": "Stable Audio 3 Medium Q8_0 GGUF",
      "default": true,
      "format": "gguf",
      "precision": "q8_0",
      "target_directory": "Stable-Audio-3-Medium-GGUF",
      "files": [
        "Stable-Audio-3-Medium-GGUF/stable-audio-3-medium-q8_0.gguf"
      ],
      "strip_prefix": "Stable-Audio-3-Medium-GGUF"
    },
    {
      "id": "stable_audio_3_medium_f16",
      "display_name": "Stable Audio 3 Medium F16 GGUF",
      "format": "gguf",
      "precision": "f16",
      "target_directory": "Stable-Audio-3-Medium-GGUF",
      "files": [
        "Stable-Audio-3-Medium-GGUF/stable-audio-3-medium-f16.gguf"
      ],
      "strip_prefix": "Stable-Audio-3-Medium-GGUF"
    },
    {
      "id": "stable_audio_3_small_music_q8_0",
      "display_name": "Stable Audio 3 Small Music Q8_0 GGUF",
      "format": "gguf",
      "precision": "q8_0",
      "target_directory": "Stable-Audio-3-Small-Music-GGUF",
      "files": [
        "Stable-Audio-3-Small-Music-GGUF/stable-audio-3-small-music-q8_0.gguf"
      ],
      "strip_prefix": "Stable-Audio-3-Small-Music-GGUF"
    },
    {
      "id": "stable_audio_3_small_music_f16",
      "display_name": "Stable Audio 3 Small Music F16 GGUF",
      "format": "gguf",
      "precision": "f16",
      "target_directory": "Stable-Audio-3-Small-Music-GGUF",
      "files": [
        "Stable-Audio-3-Small-Music-GGUF/stable-audio-3-small-music-f16.gguf"
      ],
      "strip_prefix": "Stable-Audio-3-Small-Music-GGUF"
    },
    {
      "id": "stable_audio_3_small_sfx_q8_0",
      "display_name": "Stable Audio 3 Small SFX Q8_0 GGUF",
      "format": "gguf",
      "precision": "q8_0",
      "target_directory": "Stable-Audio-3-Small-SFX-GGUF",
      "files": [
        "Stable-Audio-3-Small-SFX-GGUF/stable-audio-3-small-sfx-q8_0.gguf"
      ],
      "strip_prefix": "Stable-Audio-3-Small-SFX-GGUF"
    },
    {
      "id": "stable_audio_3_small_sfx_f16",
      "display_name": "Stable Audio 3 Small SFX F16 GGUF",
      "format": "gguf",
      "precision": "f16",
      "target_directory": "Stable-Audio-3-Small-SFX-GGUF",
      "files": [
        "Stable-Audio-3-Small-SFX-GGUF/stable-audio-3-small-sfx-f16.gguf"
      ],
      "strip_prefix": "Stable-Audio-3-Small-SFX-GGUF"
    },
    {
      "id": "stable_audio_3_medium_safetensors",
      "display_name": "Stable Audio 3 Medium Safetensors",
      "format": "safetensors",
      "precision": "native",
      "target_directory": "stable-audio-3-medium",
      "files": [
        "model_config.json",
        "model.safetensors"
      ],
      "download": {
        "kind": "huggingface_snapshot",
        "repo": "stabilityai/stable-audio-3-medium",
        "gated": true
      }
    },
    {
      "id": "stable_audio_3_small_music_safetensors",
      "display_name": "Stable Audio 3 Small Music Safetensors",
      "format": "safetensors",
      "precision": "native",
      "target_directory": "stable-audio-3-small-music",
      "files": [
        "model_config.json",
        "model.safetensors"
      ],
      "download": {
        "kind": "huggingface_snapshot",
        "repo": "stabilityai/stable-audio-3-small-music",
        "gated": true
      }
    },
    {
      "id": "stable_audio_3_small_sfx_safetensors",
      "display_name": "Stable Audio 3 Small SFX Safetensors",
      "format": "safetensors",
      "precision": "native",
      "target_directory": "stable-audio-3-small-sfx",
      "files": [
        "model_config.json",
        "model.safetensors"
      ],
      "download": {
        "kind": "huggingface_snapshot",
        "repo": "stabilityai/stable-audio-3-small-sfx",
        "gated": true
      }
    }
  ],
  "sources": [
    {
      "format": "gguf",
      "roots": {
        "model": ".",
        "weights": "$gguf"
      },
      "files": {
        "model_config": "model:model_config.json",
        "t5_config": "model:t5gemma-b-b-ul2/config.json",
        "t5_tokenizer_json": "model:t5gemma-b-b-ul2/tokenizer.json",
        "t5_tokenizer_model": "model:t5gemma-b-b-ul2/tokenizer.model",
        "t5_tokenizer_config": "model:t5gemma-b-b-ul2/tokenizer_config.json"
      },
      "tensors": {
        "model_weights": {
          "source": "weights:",
          "prefix": "model_weights"
        },
        "t5_weights": {
          "source": "weights:",
          "prefix": "t5_weights"
        }
      }
    },
    {
      "format": "safetensors",
      "roots": {
        "model": ".",
        "t5": "t5gemma-b-b-ul2"
      },
      "files": {
        "model_config": "model:model_config.json",
        "t5_config": "t5:config.json",
        "t5_tokenizer_json": "t5:tokenizer.json",
        "t5_tokenizer_model": "t5:tokenizer.model",
        "t5_tokenizer_config": "t5:tokenizer_config.json"
      },
      "tensors": {
        "model_weights": "model:model.safetensors",
        "t5_weights": "t5:model.safetensors"
      }
    }
  ]
}