Open-source CLI and platform for LLM red teaming, prompt evaluation, and application security
Core evaluation CLI and red teaming library are 100% free and open source for local and CI workflows
Promptfoo is an open-source evaluation framework and red-teaming tool designed for LLM applications. It systematically tests prompts, agent outputs, and RAG pipelines against jailbreaks, hallucinations, and security vulnerabilities. Seamlessly integrating into CI/CD pipelines, Promptfoo helps developers benchmark models, assess regressions, and ship hardened AI software with confidence.