promptfoo

Test and red-team LLM apps, prompts, and agents.

Visit site GitHub
Open source Self-hostable free
EvaluationGuardrailsData scientistPlatform engineer
In these stacks
Test and score my agents
LicenseMIT
PriceFree
UsageNo model cost
Model supportmulti
Languagestypescript
GitHub stars22.9k
Last activity2026-07-04
Verified2026-06-28
Verified byseed

A CLI and library for testing LLM outputs, running red-team security scans, comparing prompts and models, and CI-integrated eval pipelines.

Strengths

  • Red-team scanning for prompt injection and jailbreaks
  • Compares prompts and models side by side with assertion-based testing

Tradeoffs

  • Test authoring requires upfront investment
  • CLI-first; no hosted dashboard