{
  "schemaVersion": "1.0",
  "updates": [],
  "companyNews": [],
  "id": "b901",
  "language": "en",
  "name": "Moshi",
  "originalName": "Moshi",
  "library": "tools",
  "category": "Audio & music",
  "categoryKey": "音频音乐",
  "categoryKeys": [
    "音频音乐"
  ],
  "categories": [
    "Audio & music"
  ],
  "entryPoints": [],
  "summary": "Model user and assistant audio streams together for full-duplex spoken dialogue, with PyTorch, MLX and Rust inference implementations.",
  "introduction": "Moshi jointly models user audio, system audio and text, allowing listening and speaking to overlap instead of waiting for a complete recording to be transcribed. It uses the Mimi streaming codec, with PyTorch research code, Apple-device MLX, a Rust backend and web client. Hardware and quantization vary by implementation. Python and client code use MIT, Rust uses Apache, and official weights use CC-BY 4.0.",
  "capabilities": [],
  "company": {
    "name": "Kyutai",
    "originalName": "Kyutai",
    "country": {
      "codes": [
        "FR"
      ],
      "names": [
        "France"
      ],
      "display": "France",
      "basis": "headquarters",
      "basisLabel": "Company headquarters"
    }
  },
  "officialUrl": "https://github.com/kyutai-labs/moshi",
  "profileUpdatedOn": "2026-10-20",
  "sources": [
    {
      "kind": "product",
      "url": "https://github.com/kyutai-labs/moshi"
    }
  ],
  "news": [],
  "links": {
    "interactive": "products/moshi--b901.html",
    "reading": "read/en/products/moshi--b901.html",
    "markdown": "read/en/products/moshi--b901.md",
    "json": "read/en/products/moshi--b901.json"
  }
}