@langwatch/scenario
    Preparing search index...

    Interface ScenarioResult

    Represents the result of a scenario execution.

    interface ScenarioResult {
        agentTime?: number;
        audio?: VoiceRecording;
        criteria?: CriterionResult[];
        error?: string;
        evaluations?: EvaluationResult[];
        inconclusiveCriteria?: string[];
        latency?: LatencyMetrics;
        messages: ModelMessage[];
        metCriteria: string[];
        reasoning?: string;
        runId: string;
        success: boolean;
        timeline?: VoiceEvent[];
        totalTime?: number;
        unmetCriteria: string[];
    }
    Index
    agentTime? audio? criteria? error? evaluations? inconclusiveCriteria? latency? messages metCriteria reasoning? runId success timeline? totalTime? unmetCriteria
    agentTime?: number

    The time the agent spent processing during the scenario in seconds.

    audio?: VoiceRecording

    Voice-only: the full audio record of the conversation, segmented by speaker. Populated by the executor at end-of-scenario when a voice adapter participated; absent for text-only runs (back-compat).

    criteria?: CriterionResult[]

    One entry per judged criterion, in the order the scenario declared them. Absent when no judge verdict produced the result.

    error?: string

    An optional error message if the scenario failed due to an error.

    evaluations?: EvaluationResult[]

    One result per evaluator attached to the run. Present only when the scenario declared evaluators.

    inconclusiveCriteria?: string[]

    The criteria the judge could not decide from the evidence. Also present in unmetCriteria, so success semantics are unchanged; this list only separates "could not tell" from "judged false". Absent when the verdict left nothing inconclusive.

    latency?: LatencyMetrics

    Voice-only: aggregate response-time statistics across the agent's turns. Absent for text-only runs (back-compat).

    messages: ModelMessage[]

    The sequence of messages exchanged during the scenario.

    metCriteria: string[]

    A list of criteria that were successfully met.

    reasoning?: string

    The reasoning behind the scenario's outcome.

    runId: string

    Unique identifier for this scenario run.

    success: boolean

    Indicates whether the scenario was successful.

    timeline?: VoiceEvent[]

    Voice-only: timestamped events on the voice conversation timeline (user_start_speaking, agent_start_speaking, user_interrupt, etc.). Absent for text-only runs (back-compat).

    totalTime?: number

    The total time taken for the scenario execution in seconds.

    unmetCriteria: string[]

    A list of criteria that were not met.

    Settings
    Member Visibility

    On This Page
    Properties