{
  "id": "audiocraft",
  "name": "AudioCraft / MusicGen",
  "category": "music_audio",
  "description": "Explore text- and melody-conditioned music generation with Meta’s AudioCraft research toolkit and MusicGen models.",
  "useCases": [
    "Sketch instrumental musical ideas for noncommercial experimentation",
    "Study how prompts and a permitted reference melody influence generated music"
  ],
  "repoUrl": "https://github.com/facebookresearch/audiocraft",
  "docsUrl": "https://github.com/facebookresearch/audiocraft/blob/main/docs/MUSICGEN.md",
  "demo": {
    "url": "https://ai.honu.io/papers/musicgen/",
    "label": "Hear the creators’ MusicGen samples",
    "evidenceType": "creator_demo",
    "note": "Official project sample page linked by the repository, including text-conditioned and stereo examples; curated examples are not an independent quality test."
  },
  "license": {
    "code": "MIT",
    "weights": "CC-BY-NC-4.0",
    "caveat": "Code and supplied model weights have different licenses. The noncommercial weights restriction means this is not an unrestricted commercial-generation option.",
    "url": "https://github.com/facebookresearch/audiocraft/blob/main/LICENSE_weights",
    "codeUrl": "https://github.com/facebookresearch/audiocraft/blob/main/LICENSE",
    "weightsUrl": "https://github.com/facebookresearch/audiocraft/blob/main/LICENSE_weights"
  },
  "pricing": {
    "summary": "No software fee for local use within license terms; GPU compute is your cost.",
    "caveat": "The official MusicGen guide calls for a GPU and at least 16 GB of memory for medium models. Smaller models or shorter clips may fit smaller GPUs; rented hardware and hosted demo availability are separate."
  },
  "platforms": [
    "Python / PyTorch",
    "GPU workstation",
    "Local Gradio interface",
    "Official browser demo"
  ],
  "setupDifficulty": "advanced",
  "setupRequirements": [
    "Repository installation instructions specify Python 3.9 and PyTorch 2.1.0",
    "FFmpeg is recommended; check the pinned dependency stack before choosing an environment",
    "Download model checkpoints; allow GPU memory and runtime for the chosen model and clip duration"
  ],
  "limitations": [
    "MusicGen does not generate realistic vocals",
    "English prompts are best supported; styles and cultures are represented unevenly",
    "Outputs can collapse to silence, and prompt experimentation may be necessary",
    "A research toolkit with noncommercial model weights needs additional review before downstream deployment"
  ],
  "evidenceCheckedDate": "2026-10-14",
  "testStatus": "not_tested",
  "sources": [
    {
      "title": "Official AudioCraft README: setup and model scope",
      "url": "https://github.com/facebookresearch/audiocraft",
      "evidenceCheckedDate": "2026-10-14"
    },
    {
      "title": "Actual code LICENSE: MIT",
      "url": "https://github.com/facebookresearch/audiocraft/blob/main/LICENSE",
      "evidenceCheckedDate": "2026-10-14"
    },
    {
      "title": "Actual weights LICENSE: CC BY-NC 4.0",
      "url": "https://github.com/facebookresearch/audiocraft/blob/main/LICENSE_weights",
      "evidenceCheckedDate": "2026-10-14"
    },
    {
      "title": "MusicGen documentation: hardware and demos",
      "url": "https://github.com/facebookresearch/audiocraft/blob/main/docs/MUSICGEN.md",
      "evidenceCheckedDate": "2026-10-14"
    },
    {
      "title": "MusicGen model card: limitations",
      "url": "https://github.com/facebookresearch/audiocraft/blob/main/model_cards/MUSICGEN_MODEL_CARD.md",
      "evidenceCheckedDate": "2026-10-14"
    },
    {
      "title": "Creators’ MusicGen sample page",
      "url": "https://ai.honu.io/papers/musicgen/",
      "evidenceCheckedDate": "2026-10-14"
    }
  ],
  "categoryLabel": "Music & Audio",
  "tags": [
    "music generation",
    "MusicGen",
    "melody",
    "research",
    "noncommercial weights"
  ],
  "card": {
    "cost": "GPU compute; noncommercial weights",
    "setup": "Python + GPU setup"
  },
  "setup": {
    "summary": "Python + GPU setup",
    "requirements": [
      "Repository installation instructions specify Python 3.9 and PyTorch 2.1.0",
      "FFmpeg is recommended; check the pinned dependency stack before choosing an environment",
      "Download model checkpoints; allow GPU memory and runtime for the chosen model and clip duration"
    ]
  },
  "evidence": {
    "status": "documentation_reviewed",
    "demoType": "creator_demo",
    "independentlyTested": false,
    "note": "Official source documentation reviewed; no repository installed or executed."
  },
  "releaseDate": null,
  "releaseDateNote": "Not researched; the evidence check date does not indicate a recent release."
}
