Checks AI-generated text for hallucinations, PII, toxicity, bias, and prompt injection in one call.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.