Evaluates and scores how faithfully humor is preserved in translated text, providing fidelity metrics and explanatory feedback.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.