An AI evaluation and safety platform that automatically scores an LLM's outputs against test cases for correctness, safety and hallucination, aimed at testing an application before and after deployment.