Prompt playground

Pick an AI feature, edit its prompt, and see how prompt quality moves the score — then gate a change against your baseline.

Mode

Simulated, offline, and free — answers are graded deterministically to show the flow.

Guided demo

One click each — a strong-prompt baseline, then a gate with the same prompt (passes) or a weak prompt (fails).

Run history

No runs yet. Run an eval to populate history.

Pick a feature to begin

Choose an AI feature from the dropdown, then Run eval on its prompt. Or hit the one-click Passing / Failing gate demo.