Understand what a run costs
See the estimate for a run, where the cost comes from, and what happens when an agent reaches its monthly budget.
Every model call an agent makes costs something, so every run has a cost. You can see it before you scale and control it with budgets.
Where cost comes from
A run is a loop of model calls. The cost follows how many turns it takes, how much text goes in and out (tokens), and which model it uses. A stronger model costs more per call than a faster one.
See the cost
- On a run. The run header shows the steps, tokens and cost.
- Before you run. In the Budget panel on the agent page, a line such as "About ... per run" shows the typical cost, with a range. It is based on recent runs, or on model prices if there are none yet. It is an estimate, not a promise.
- This month. The same panel shows how much of the monthly budget has been used, with a bar that changes colour as it fills.
- Across agents. Usage shows tokens and cost per day and per agent.
Figures are estimates from token counts. Your AI provider's invoice, or your account's billing page if the run used managed AI, is the true amount. If your account has a flat fee per run, it is shown separately from the AI cost, and it does not count against the budget.
Budgets
- Per run (USD) is a hard stop. A run that reaches it ends with a plain message.
- Per month (USD, 0 = no limit) is a monthly ceiling.
You also get notifications as an agent's month fills up.
When the monthly budget is reached
The run does not fail silently. It pauses, and an approval card titled " reached its monthly budget" appears in Approvals:
- Allow this run lets just this run finish.
- Raise this month by adds the amount you enter to this month's budget only.
- Decline ends the run as Cancelled.
A task that is waiting shows Waiting for your budget decision. Test runs and replays still fail at the budget instead of asking.
Bring cost down
- Choose a lighter Model where the quality is enough.
- Give clearer instructions so the agent needs fewer turns.
- Fetch less. Pages and results that are shorter cost less.
- Run less often.
More guides in PGB Agents
What a PGB Agent is and what it can do
An agent is a saved job with instructions, a list of allowed tools, a budget and a record of every step it takes.
Create an agent from scratch or from the Marketplace
Make a blank agent in two clicks, or add a ready-made one from PGB Marketplace and review it before it runs.
Write good instructions for an agent
How to word the Instructions box so an agent does the right job, stays inside its limits and tells you when it is unsure.
Choose tools and understand which need approval
The list of tools an agent can use, what each does, and which ones always pause for your approval.
Start agents manually, on a schedule or from a webhook
Run an agent by hand, set it to repeat with the schedule builder, or give it a private web address that starts a run.
Set budgets and know when a run will stop
Put a dollar limit on each run and each month, and learn the other limits that stop an agent from running away.