Behavioral call-replay for a check's model change (defaults to mock — no token spend)
Re-issue a representative recorded model call per affected trace against the candidate model and diff vs the recorded answer.
mode=mock (default) never spends tokens — safe for a dashboard click.
mode=real performs live same-provider calls with the server’s key
(opt-in; costs tokens). Same-provider MVP (OpenAI/Gemini).
Authorizations
Enter your API key (e.g. dai_sk_test_key_001)
Headers
Path Parameters
Query Parameters
mock (default, no spend) returns deterministic outputs; real performs live same-provider calls with the server key.
mock, real Which recorded call(s) per affected trace to replay. first = the representative (earliest) call; all is reserved.
first, all Also run the LLM-judge tier on changed samples (better/worse/equivalent vs the recorded answer). In real mode this spends a judge call per changed sample.
Cookies
Response
Successful Response