Optional createCreates a new tracker for this AI Config invocation. Each call returns a new tracker with a fresh runId. Use createTracker() at the start of each execution to obtain a tracker, then use it to record metrics for that run.
Whether the configuration is enabled.
Optional evaluationEvaluation metric key for judge configurations. The key of the metric that this judge can evaluate.
Optional evaluationEvaluation metric keys for judge configurations (legacy). The keys of the metrics that this judge can evaluate.
Use evaluationMetricKey instead. This field is kept for legacy support.
The key of the AI Config.
Optional messagesOptional prompt data for judge configurations.
Optional modelOptional model configuration.
Optional providerOptional configuration for the provider.
Generated using TypeDoc
Judge-specific AI Config with required evaluation metric key.