Use case
How to cap one AI agent's spend at a fixed amount per month
Example: a monthly limit for one agent, without touching any other agent's budget, and a clear choice between monitoring and strict enforcement.
Create one budget scoped to that agent's agent_id with a monthly limit. With strict mode on, calls whose estimated cost would exceed it are refused. With strict mode off, the budget records spend and sends threshold alerts without refusing anything.
How it works
Budgets can be set at org, team, agent and customer level. An agent-scoped budget counts only spend attributed to that agent_id and resets each UTC calendar month. A team or org budget can apply at the same time, and each strict budget is checked independently before forwarding. Whether strict or not, each configured budget sends threshold alerts at 50, 80, 90, 100 and 120% to a Slack or Discord webhook you set up.
{
"scope_type": "agent",
"scope_id": "research-agent",
"monthly_limit_usd": 300,
"strict_mode": true
}What to know before relying on it
Alerts go only to Slack or Discord webhooks; there is no email delivery. A strict budget decides from an estimate, so a single long response can settle above the limit.
Why it matters
A single company-wide budget cannot answer "did this agent overspend?" A budget scoped to the agent isolates the problem to its source.
Frequently asked questions
What happens on the first day of the next month?
The monthly scope resets and spend starts accumulating from zero again for that agent.
Can I set this to alert instead of block?
Yes. Leave strict_mode off and the budget records spend and sends threshold alerts without refusing calls. Turn it on to refuse calls whose estimated cost would exceed the limit. Choose deliberately when you create the budget: on the production service new budgets are strict unless you set otherwise.
See what this would look like in your product.
Start with a guided 30-day free pilot on live production traffic. The 30 days start when your traffic flows through Spendline, not at sign-up. We agree the scope, success criteria and paid terms before you start. There is no Spendline subscription fee during the pilot; your model-provider charges still apply.
Prefer email? fida@spendline.ai