{
  "protocolVersion": "0.3.0",
  "name": "Anyray",
  "description": "Self-hosted AI gateway that cuts LLM-inference spend with real-time call optimization, spend visibility and governance, and full prompt/response privacy. Exposes an OpenAI-compatible API in front of OpenAI, Anthropic, Vertex AI, Bedrock, and Azure, and runs entirely in the customer's own environment.",
  "url": "https://docs.anyray.ai/developers/api-reference",
  "preferredTransport": "HTTP+JSON",
  "provider": {
    "organization": "Anyray",
    "url": "https://anyray.ai"
  },
  "version": "1.0.0",
  "documentationUrl": "https://docs.anyray.ai",
  "iconUrl": "https://anyray.ai/anyray_logo.svg",
  "capabilities": {
    "streaming": true,
    "pushNotifications": false,
    "stateTransitionHistory": false
  },
  "defaultInputModes": ["application/json", "text/plain"],
  "defaultOutputModes": ["application/json", "text/plain"],
  "skills": [
    {
      "id": "chat-completions",
      "name": "Optimized chat completions",
      "description": "Proxy an OpenAI-compatible chat-completions request to the configured provider, applying real-time optimization (prompt and context compression, tool pruning, semantic cache) before forwarding. Supports streaming responses.",
      "tags": ["llm", "openai-compatible", "gateway", "cost-optimization"],
      "examples": [
        "POST /v1/chat/completions with an OpenAI-format body to any routed provider model."
      ]
    },
    {
      "id": "embeddings",
      "name": "Embeddings",
      "description": "Proxy an OpenAI-compatible embeddings request to the configured provider through the gateway.",
      "tags": ["llm", "openai-compatible", "embeddings"],
      "examples": ["POST /v1/embeddings with an OpenAI-format body."]
    },
    {
      "id": "spend-attribution",
      "name": "Spend attribution and governance",
      "description": "Every request is attributed (user, team, model, provider, tokens, cost, latency, decision) into a content-free spend store for per-user and per-team governance.",
      "tags": ["observability", "cost", "governance"],
      "examples": ["Inspect attributed spend per user and team in the self-hosted console."]
    }
  ],
  "additionalInterfaces": [
    { "transport": "HTTP+JSON", "url": "https://docs.anyray.ai/developers/api-reference" }
  ],
  "securitySchemes": {
    "apiKey": {
      "type": "http",
      "scheme": "bearer",
      "description": "Bearer API key issued by your self-hosted Anyray deployment. See https://docs.anyray.ai for authentication."
    }
  },
  "deployment": "self-hosted",
  "openapi": "https://docs.anyray.ai/openapi.json",
  "llmsTxt": "https://anyray.ai/llms.txt",
  "llmsFullTxt": "https://anyray.ai/llms-full.txt",
  "contact": { "email": "hi@anyray.ai" }
}