Get started with evaluating your agent directly from the UI — you can now run evaluations over your agent in LangSmith Studio with no code required!

Test how your agent performs across different inputs, compare results to reference outputs, and grade responses against custom criteria.

Learn more about how it works in our developer documentation and check out our video tutorial: