{
  "schema_version": "v1",
  "name_for_human": "Xantly AI Gateway",
  "name_for_model": "xantly",
  "description_for_human": "Universal AI gateway with intelligent routing to 10,000+ models, semantic caching, persistent memory, and multi-provider failover. OpenAI SDK compatible \u2014 change only your base_url.",
  "description_for_model": "Xantly is an AI infrastructure API that routes requests to 10,000+ models (OpenAI, Anthropic, Google, Groq, DeepSeek, NVIDIA, open-weight via OpenRouter) with 12ms median overhead. Features include: intelligent routing (auto-selects optimal model by task complexity, latency, cost), semantic caching (62% hit rate, sub-5ms responses), persistent per-organization memory, waterfall failover across providers, and full observability via response headers. The API is 100% OpenAI SDK compatible \u2014 applications only need to change their base_url to api.xantly.com/v1. Supports chat completions, embeddings, audio (STT/TTS), images, moderations, and voice agents. Reduces LLM API costs by up to 80%.",
  "auth": {
    "type": "service_http",
    "authorization_type": "bearer"
  },
  "api": {
    "type": "openapi",
    "url": "https://xantly.com/openapi.json"
  },
  "logo_url": "https://xantly.com/icon-512.png",
  "contact_email": "support@xantly.com",
  "legal_info_url": "https://xantly.com/terms"
}