@conciergus/chat - v0.3.1
    Preparing search index...

    Interface MessageMetadata

    Comprehensive message metadata interface Supports all AI SDK 5 metadata fields plus enterprise extensions

    interface MessageMetadata {
        duration?: number;
        model?: string;
        finishReason?:
            | "cancelled"
            | "stop"
            | "length"
            | "content-filter"
            | "tool-calls"
            | "function-call";
        totalTokens?: number;
        inputTokens?: number;
        outputTokens?: number;
        cost?: number;
        provider?: string;
        modelVersion?: string;
        endpoint?: string;
        timestamp?: string
        | Date;
        firstTokenTime?: number;
        averageTokenTime?: number;
        confidence?: number;
        perplexity?: number;
        relevance?: number;
        [key: string]: any;
    }

    Indexable

    • [key: string]: any

      Custom metadata fields

    Index

    Properties

    duration? model? finishReason? totalTokens? inputTokens? outputTokens? cost? provider? modelVersion? endpoint? timestamp? firstTokenTime? averageTokenTime? confidence? perplexity? relevance?

    Properties

    duration?: number

    Response generation duration in milliseconds

    model?: string

    AI model used for generation

    finishReason?:
        | "cancelled"
        | "stop"
        | "length"
        | "content-filter"
        | "tool-calls"
        | "function-call"

    Finish reason for the response

    totalTokens?: number

    Total tokens used in the conversation

    inputTokens?: number

    Input/prompt tokens

    outputTokens?: number

    Output/completion tokens

    cost?: number

    Estimated cost in USD

    provider?: string

    AI provider (openai, anthropic, etc.)

    modelVersion?: string

    Provider-specific model version

    endpoint?: string

    API endpoint used

    timestamp?: string | Date

    Request timestamp

    firstTokenTime?: number

    Time to first token

    averageTokenTime?: number

    Average time between tokens

    confidence?: number

    Response confidence score (0-1)

    perplexity?: number

    Perplexity score for language models

    relevance?: number

    Relevance score for RAG responses

    Settings

    Member Visibility

    On This Page

    Properties