Audits and ranks LLM outputs for accuracy, completeness, and format adherence.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.