Software product / application
Promptfoo
Local-first testing and red teaming for LLM applications, with assertions, provider comparisons and browser-based results.
About this listing
Promptfoo helps developers compare prompts, models and RAG configurations against repeatable test cases. Its CLI and library run evaluations with configurable assertions, while a browser results view makes it easier to inspect output differences. Caching and concurrency support repeated evaluation during development.
It also supports adversarial testing of LLM applications and can fit into CI/CD workflows. The evaluation runner is local-first, but configured model providers and optional hosted features can receive data; local execution alone is not a guarantee of an offline deployment.
The Community offering includes core evaluation features and a limited red-teaming allowance. Enterprise plans add team controls and managed capabilities with custom pricing. Separately selected model APIs and infrastructure can incur their own costs.