ART Framework API Docs
    Preparing search index...

    Interface ExecutionMetadata

    Metadata summarizing an agent execution cycle, including performance metrics and outcomes.

    ExecutionMetadata

    interface ExecutionMetadata {
        error?: string;
        llmCalls: number;
        llmCost?: number;
        llmMetadata?: LLMMetadata;
        status: "success" | "error" | "partial";
        threadId: string;
        toolCalls: number;
        totalDurationMs: number;
        traceId?: string;
        userId?: string;
    }
    Index

    Properties

    error? llmCalls llmCost? llmMetadata? status threadId toolCalls totalDurationMs traceId? userId?

    Properties

    error?: string

    A top-level error message if the overall status is 'error' or 'partial'.

    llmCalls: number

    The number of calls made to the ReasoningEngine.

    llmCost?: number

    An optional estimated cost for the LLM calls made during this execution.

    llmMetadata?: LLMMetadata

    Aggregated metadata from LLM calls made during the execution.

    status: "success" | "error" | "partial"

    The overall status of the execution ('success', 'error', or 'partial' if some steps failed but a response was generated).

    threadId: string

    The thread ID associated with this execution cycle.

    toolCalls: number

    The number of tool execution attempts made by the ToolSystem.

    totalDurationMs: number

    The total duration of the agent.process() call in milliseconds.

    traceId?: string

    The trace ID used during this execution, if provided.

    userId?: string

    The user ID associated with the execution, if provided.

    Settings

    Member Visibility

    On This Page

    Properties