{
  "protocolVersion": "0.3.0",
  "name": "Rauta Advisor",
  "description": "Recommends GPU cloud infrastructure grounded in live, measured data: market prices, VRAM-by-precision model sizing, and benchmarker-sourced cold-start and reliability telemetry.",
  "url": "https://mcp.rauta.ai/mcp",
  "preferredTransport": "JSONRPC",
  "provider": {
    "organization": "Rauta",
    "url": "https://rauta.ai"
  },
  "version": "0.1.0",
  "documentationUrl": "https://rauta.ai/install",
  "capabilities": {
    "streaming": false,
    "pushNotifications": false,
    "stateTransitionHistory": false
  },
  "defaultInputModes": ["text/plain"],
  "defaultOutputModes": ["text/plain", "application/json"],
  "skills": [
    {
      "id": "recommend-gpu-infra",
      "name": "Recommend GPU infrastructure",
      "description": "Given a model or workload, return a price-ranked shortlist of real provider+GPU listings that fit the VRAM requirement, with reliability and cold-start caveats. Never invents prices or latencies.",
      "tags": ["gpu", "infrastructure", "llm", "inference", "cost", "reliability"],
      "examples": [
        "What is the cheapest GPU to serve Llama-3.3-70B at int4?",
        "Is Lambda Labs reliable for provisioning this week?",
        "How much VRAM does Qwen2.5-14B need and what fits it?"
      ]
    }
  ]
}