Which answer is better, and how bad is the worse one?

Paste two AI answers to the same question. Each gets a P0-P3 severity, a reason and a suggested fix.

Question / prompt

Answer A Answer B

Rubric weights

Bias control: judge in both orders (A,B and B,A) and flag disagreement

Judge engine

Engine

ProviderModel

Your API key (sent only with this request to the judge API, never stored) Reference material (optional, makes claim checks grounded)

Batch mode

Upload a CSV with columns: prompt,answer_a,answer_b. Uses the weights and engine from the Judge tab.