Tool profile
Promptfoo
EvaluationA command-line tool and library for configurable evaluations of prompts, models, agents and retrieval applications.
Open sourceCode / delivery model
LocalWhere it runs
Not ratedNo hands-on score claimed
Where it may fit.
Turning a small task set and explicit assertions into repeatable evaluation runs.
Before you adopt it.
- Model-provider calls may send data off-device and incur separate charges.
- Assertions and model-based graders need their own validity checks.
- Run security tests only against systems you are authorized to assess.
Verify the current details.
Versions, pricing, licenses, and availability can change. Use the original project to confirm what applies to your use case.