@squidcloud/client
    Preparing search index...

    Interface KnowledgeBaseGraphUsage

    Aggregated LLM token usage (and best-effort cost) of graph work on a KB. Token counts are the provider-reported ground truth; estCostUsd is estimated from model list prices at call time — recompute from the tokens for exact accounting.

    interface KnowledgeBaseGraphUsage {
        byModel?: KnowledgeBaseGraphModelUsage[];
        cacheReadTokens: number;
        cacheWriteTokens: number;
        estCostUsd: number;
        inputTokens: number;
        llmCalls: number;
        outputTokens: number;
        reasoningTokens: number;
    }
    Index

    Properties

    byModel? cacheReadTokens cacheWriteTokens estCostUsd inputTokens llmCalls outputTokens reasoningTokens

    Properties

    byModel?: KnowledgeBaseGraphModelUsage[]

    Per-model split of the totals above (LLM calls only), present when any usage was recorded. The per-model figures sum to the aggregate; embedding usage is excluded from both.

    cacheReadTokens: number

    Input tokens served from the provider's prompt cache.

    cacheWriteTokens: number

    Input tokens written into the provider's prompt cache.

    estCostUsd: number

    Best-effort USD estimated from list prices at call time.

    inputTokens: number

    Prompt (input) tokens.

    llmCalls: number

    Number of LLM calls.

    outputTokens: number

    Completion (output) tokens.

    reasoningTokens: number

    Reasoning tokens, where the provider reports them separately.

    Settings

    Member Visibility

    On This Page

    Properties