Promptfoo is an open-source framework for evaluating, testing, and red-teaming LLM applications that helps AI developers systematically measure prompt quality, detect regressions, and identify safety vulnerabilities before deployment. It supports automated testing across multiple LLM providers simultaneously, comparison of different model versions, and adversarial probing for jailbreaks and harmful outputs. AI engineering teams use Promptfoo as their CI/CD layer for prompt engineering — ensuring that every change to prompts or models is rigorously evaluated before reaching production users.