Interface: TokenUsage¶

Defined in: src/types.ts:194

LLM token usage totals, as accumulated by kaboo-workflows from each agent completion. Appears on a StreamGroup (per-agent totals) and in ActivityState.usageByRun (per-run rollup). Counts accumulate across repeat completions (interrupt resumes, delegate re-entry).

Properties¶

cacheReadInputTokens?¶

optional cacheReadInputTokens?: number

Defined in: src/types.ts:206

Input tokens served from the provider's prompt cache (billed at a steep discount), counted apart from TokenUsage.inputTokens. Absent on streams from services that predate cache accounting.

cacheWriteInputTokens?¶

optional cacheWriteInputTokens?: number

Defined in: src/types.ts:208

Input tokens written to the provider's prompt cache. Absent on older streams.

cost?¶

optional cost?: number

Defined in: src/types.ts:214

Accumulated dollar cost of the LLM calls, as reported by the provider's usage accounting. Absent on older streams or providers without per-call cost reporting.

inputTokens¶

inputTokens: number

Defined in: src/types.ts:196

Prompt/input tokens consumed, not counting cache reads or writes.

outputTokens¶

outputTokens: number

Defined in: src/types.ts:198

Completion/output tokens produced.

totalTokens¶

totalTokens: number

Defined in: src/types.ts:200

Total tokens (as reported by the model provider).