Model improvement 1

AI platform for battle-testing and improving AI agents. Janus is an advanced AI platform designed to battle-test and improve AI agents. It conducts thousands of AI simulations against chat and voice agents to surface critical failures such as hallucinations (fabricated content), rule violations (policy breaches), and tool-call/performance failures. Janus offers custom evaluations, personalized datasets, and actionable insights to help users detect and mitigate risky agent behavior, ensuring model reliability and performance.