Human Judgment Crowdsourcing via x402 Micropayment

Choose this endpoint when you need genuine human judgment on subjective, nuanced, or culturally specific questions that LLMs cannot reliably answer — such as native-speaker naturalness checks, content moderation edge cases, or translation quality ratings. Prefer it over LLM-only approaches when human consensus is required for trust, compliance, or QA pipelines. It is best suited for asynchronous batch workflows where a short wait for human responses is acceptable.

• Missing required 'question' parameter returns an error
• Invalid 'responses' value above max of 5 rejected
• Reward value outside 0.02–0.05 range is clamped or rejected
• Payment failure via x402 USDC prevents task from being submitted
• Slow fill rate if reward is set too low or question is complex
• No consensus reached if human answers are evenly split across labels