Grades LLM responses for relevance, completeness, clarity, confidence, and overall quality.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.