Agent usage and limits
Every agent run uses AI tokens. This page shows where to see how much each agent uses, how per-agent limits work, and what Helpin checks before a run starts.
See an agent's usage
Go to Automation > Agents and select the agent.
Open a tab:
Analytics — run trends and the last five runs, including the tokens each one used.
Limits — Monthly limit, Used this month and Total tokens across all recorded runs.
Usage on the Limits tab covers all versions of the agent. If Total tokens says "At least this many tokens used", the figure is a minimum and the true total may be higher.
Set a monthly token limit
On the Limits tab, under Monthly token limit:
Choose one of the preset limits, or select Custom and enter a number of tokens.
Select Apply.
A progress bar then shows how much of the limit the agent has used this month. It turns amber and then red as the agent gets close to the limit. To remove the limit, choose **No limit**.
Use the monthly limit as a tracking target. Today it doesn't stop every kind of run when an agent goes over it, and it isn't yet enforced for custom agents. For firm control over spending, use your workspace's AI usage settings.
What Helpin checks before a run
Before any agent run starts, Helpin checks that the run is allowed:
Access — you can use the agent, and the agent can work on the target you picked.
Plan features — on Helpin Cloud, some capabilities, such as custom agents and scheduled flows, are available on plans that include them. If your plan doesn't include one, Helpin shows a message instead of starting the run.
AI usage — on Helpin Cloud, the run is checked against your workspace's AI usage for the AI model it will use. If the workspace has used its included AI usage and extra AI usage is turned off, the run doesn't start. A run can also stop before it would go over what's available.
If the workspace is locked because there's no active plan, runs don't start until an admin chooses a plan.
Manage workspace AI usage on Helpin Cloud
Workspace admins manage AI usage in Settings > Billing. There you can see Included AI usage for the current billing period and turn Extra AI usage on or off. Included usage resets each billing period.
Self-hosted Helpin Community
On self-hosted Community, Helpin records each run's token usage so you can see it on the Analytics and Limits tabs, but Helpin doesn't charge for it. Your AI provider bills its own account for the models you connect.
Keep usage predictable
Choose an AI profile that fits the job. Smaller models suit routine work. See Agent settings reference.
Give agents focused instructions and only the tools they need, so runs finish in fewer steps.
Check Analytics after changing a version, and compare token use with earlier runs.
Related
Was this article helpful?