Promptfoo is an open-source evaluation and AI-security toolkit for LLM applications, agents and RAG systems. It lets teams define prompts, providers, test cases and assertions in configuration, then run repeatable evaluations locally, in CI or through a web review workflow instead of relying on manual prompt checks.
The current official positioning is broader than prompt regression testing. Promptfoo now says it is part of OpenAI and highlights Red Teaming, Guardrails, Model Security, MCP Proxy, Code Scanning and Evaluations. That makes it relevant for security teams reviewing jailbreaks, unsafe tool use, prompt injection, model-risk gaps and MCP-mediated agent workflows.
Promptfoo works best as the evaluation and AI-security layer of an LLMOps stack. It can gate prompt and model changes before deployment, compare providers, and run adversarial tests, but teams may still need separate observability, tracing, production feedback and incident-response systems for live operations.
