three.ws Audio — MCP Server 安装与配置

{
  "id": "io.github.nirholas/audio-mcp",
  "type": "mcp-server",
  "version": "0.1.0",
  "displayName": "three.ws Audio",
  "description": "Text-to-speech, speech-to-text, audio-to-face lipsync, and motion-capture clips for 3D agents.",
  "repository": {
    "url": "https://github.com/nirholas/three.ws",
    "source": "github",
    "subfolder": "packages/audio-mcp"
  },
  "homepage": "https://three.ws",
  "distribution": {
    "packages": [
      {
        "registryType": "npm",
        "identifier": "@three-ws/audio-mcp",
        "version": "0.1.0",
        "runtimeHint": "npx",
        "transport": "stdio",
        "environmentVariables": [
          {
            "name": "THREE_WS_BASE",
            "description": "Base URL of the three.ws API that serves the audio/animation endpoints. Override only when self-hosting or targeting a preview deployment."
          },
          {
            "name": "THREE_WS_API_KEY",
            "description": "Optional three.ws API key (bearer token). Raises the metered TTS/ASR/Audio2Face rate limits and unlocks your own private/unlisted motion-capture clips. Leave unset to use the public surface anonymously.",
            "isSecret": true
          },
          {
            "name": "THREE_WS_TIMEOUT_MS",
            "description": "Per-request timeout in milliseconds for the live audio/animation calls (synthesis and recognition are real upstream model calls)."
          }
        ]
      }
    ],
    "remotes": []
  },
  "dependencies": [],
  "installTargets": [
    "claude-code",
    "claude-desktop",
    "cursor",
    "vscode"
  ],
  "keywords": [],
  "provenance": {
    "origin": "official-mcp-registry",
    "originalId": "io.github.nirholas/audio-mcp",
    "originalUrl": "https://registry.modelcontextprotocol.io/v0.1/servers/io.github.nirholas%2Faudio-mcp/versions/latest",
    "isOfficial": true,
    "status": "active"
  }
}