{
  "name": "AI Eval API",
  "description": "Automated evaluation of LLM responses — accuracy, relevance, safety scores.",
  "version": "1.0",
  "endpoints": [
    {
      "path": "/ai-eval/score",
      "method": "POST",
      "body": {
        "prompt": "string",
        "response": "string",
        "criteria": "string[]?"
      },
      "desc": "Score AI response quality"
    },
    {
      "path": "/ai-eval/compare",
      "method": "POST",
      "body": {
        "prompt": "string",
        "responses": "string[]"
      },
      "desc": "Compare multiple responses"
    },
    {
      "path": "/ai-eval/metrics",
      "method": "POST",
      "body": {
        "text": "string"
      },
      "desc": "Text quality metrics (length, readability, tokens)"
    }
  ]
}