Overview
Promptfoo is an open-source CLI and library for evaluating and red-teaming LLM applications. It facilitates test-driven LLM development by enabling users to build reliable prompts, models, and RAGs with benchmarks tailored to their use-case. The tool supports automated red teaming and pentesting to secure apps and speeds up evaluations through caching, concurrency, and live reloading. Users can score outputs automatically by defining metrics. It integrates with various LLM APIs including OpenAI, Anthropic, Azure, Google, HuggingFace, and open-source models like Llama. Promptfoo provides continuous monitoring, centralized security dashboards, and actionable remediation steps to close the loop between finding and fixing vulnerabilities. Its architecture supports integration into CI/CD pipelines, GitHub, GitLab, and Jenkins and can be deployed on-premise or in the cloud.
Common tasks