Feature · Platform

Usage & spend

See worker time, tokens, tools and model costs. Set spend limits.

Last changed

Generated from commit be57775 ·

How it works

The limits

  • Organization daily budget: what every worker together can spend in one UTC day. Set it in the Defaults card on Settings. $0 stops all model spend.
  • Worker daily budget and per-item budget: on the worker's Governance settings. 0 blocks all spend for that worker.
  • Item budget: a work item can carry its own cap.
  • Model account budgets: daily and monthly limits per account. See Model accounts and API keys.
  • Ask above: one spend above this amount asks you first. See The Always ask list.

A worker's limits are also bounded by every manager above it, unless limits.inherit is "Own limits only".

Defaults

  • A new organization from sign-up starts with the sign-up daily budget that your Agent server sets. By default it is $2 a day.
  • Other organizations use the platform default, $5 a day by default.
  • The recommended setup can suggest another value. See Recommended setup.

How it works

Agent reserves the spend before each call. When a limit cannot hold the call, Agent refuses it. When Agent cannot read the budget records, it refuses the spend, to be safe.

See where the money went

Budgets & Spend shows spend per day, by provider and model, top spenders, by purpose and per task, over 7, 30 or 90 days. The meters say Within budget, Near the limit or Limit reached: new spend is refused.

When a limit is reached

From the help article Budgets and spend limits, checked against the product on .

What you can do

The operations this feature's screens call. Each one is POST /api/<name>, and the badge says who may call it.

  • getSpendSummary

    Viewer

    Gets the spend summary.

    Types: GetSpendSummaryRequest returns SpendSummaryResponse

  • getSpendSavings

    Viewer

    Gets spend savings.

    Types: GetSpendSavingsRequest returns SpendSavingsResponse

  • getOrgSettings

    Viewer

    Gets org settings.

    Types: EmptyRequest returns OrgSettingsDto

  • setEffortBaseline

    Admin

    Sets the effort baseline.

    Types: SetEffortBaselineRequest returns EffortBaselinesResponse

  • listEffortBaselines

    Viewer

    Lists effort baselines.

    Types: EmptyRequest returns EffortBaselinesResponse

  • recordMyMeasuredEffort

    ViewerNo API tokens

    Records a my measured effort.

    Types: RecordMyMeasuredEffortRequest returns RecordMyMeasuredEffortResponse

  • deleteMyMeasuredEffort

    ViewerNo API tokens

    Deletes the my measured effort.

    Types: EmptyRequest returns EffortBaselinesResponse

  • getLiveSpend

    Viewer

    Gets live spend.

    Types: GetLiveSpendRequest returns LiveSpendResponse

  • getTaskMetrics

    Viewer

    Gets task metrics.

    Types: GetTaskMetricsRequest returns TaskMetricsResponse

  • getWorkerInsights

    Viewer

    Gets worker insights.

    Types: GetWorkerInsightsRequest returns WorkerInsightsResponse

  • listWorkItemDelegates

    Viewer

    Lists work item delegates.

    Types: ListWorkItemDelegatesRequest returns WorkItemDelegatesDto

  • listMembers

    Viewer

    Lists members.

    Types: EmptyRequest returns ListMembersResponse

Screens

What it records

The feature writes these events to the audit log, the permanent record of what happened.

  • Policy changed
  • Budget reserved
  • Checkpoint saved
  • Context retrieved
  • Delegate ended
  • Effort baseline changed
  • Integration changed
  • Measured effort recorded
  • Measured effort removed
  • Memory recalled
  • Model called
  • Policy evaluated

Generated from code at commit be57775 ·

Turn the numbers into shipped work.

Agent runs these choices for you: a persistent AI worker with memory and rules, on your Claude and Codex subscriptions.