Costs & Budgets
Set the budget, assign its scope, and choose the action at the limit. See usage as it happens, with the same controls for people and agents.
Talk to Sales about your 30-day free trial
Before the next model call
AI Gateway checks model access, rate limits and the applicable budget.
One budget
An explicit next step
Apply limits by organization, team, user, or model. The runtime checks applicable budgets before the model call and follows your configured policy.
- 1
Allow an overage
Continue the call and record usage above the cap.
- 2
Alert the owner
Notify the budget owner while calls continue.
- 3
Hold for approval
Pause further calls for human approval.
- 4
Block new calls
Stop until the budget is raised or resets.
Follow spend
back to the work
Request-level attribution connects cost to a person, team, department, cost center, agent, model, and provider. Review totals without losing the context behind them.
Runtime metering supports control and allocation. Provider invoices remain the billing record.
Plan your AI cost controlsSpend and security
In the same console
Read cost alongside adoption and security activity. Export attributed usage for finance reviews, and trace governed interactions back to the decisions recorded by AI Security Runtime.
Monitor
- Spend to date
- Remaining budget
- Usage trends
Attribute
- People and agents
- Teams and cost centers
- Models and providers
Review
- Chargeback and showback
- Budget-owner decisions
- Finance reporting
- Observe
- Detect
- Enforce
- Trace