Select a scenario, hit Analyze, and watch LLMGuard identify the exact words, timestamps, and risk scores of every hallucination — in under 20ms.