← All use cases

Evaluate

Human judgment on your content. A/B comparisons, safety review, moderation, labeling, feedback.

Batch upload (CSV)

One comparison per row: prompt, response_a, response_b. First row is the header. Up to 10,000 rows.

Workers pick one. Use any labels: safe/unsafe, 1-5, approve/reject, violation categories, etc.

$

Total: $0.80

1 item × 3 evaluators × $0.25 + 20% fee