About This Launch
Eval-only dataset for HVAC/plumbing intake and after-hours triage agents. Includes labeled cases, promptfoo harness, deterministic scorers, FAIL-bait cases. Unzip, swap provider, read the score table. Not a phone bot.
Why I Built This
We built this to make prompt changes measurable before they reach real HVAC and plumbing intake workflows. The labeled cases and deterministic scorers help teams catch failures quickly instead of relying on demos or intuition.
Launch Details
CategoryAI / ML
StageLive
ListedSeptember 8, 2026