Skip to content

AI agent action limits

All plans Member

When an AI agent works in Onplana (the in-app assistant, or a connected MCP client like Claude, Cursor, or ChatGPT), each change it makes is an action. On the Free and Starter plans, the number of times an agent can perform a given write action per month is capped. On Pro and above, agent actions are unlimited.

This keeps automated activity reasonable on the entry plans without affecting normal work: you can always create tasks, comments, and the rest yourself in the app with no cap. The limit applies only to actions an AI agent performs.

An action is one write tool call made by an agent, such as creating a task, adding subtasks, or moving a task to a sprint. Each tool is counted on its own, per calendar month.

Reads are never capped. Listing, getting, and searching work on every plan with no monthly limit, so an agent can always look things up before it acts.

Caps apply on Free and Starter only. On Pro, Business, Enterprise, and Enterprise Plus, agent actions are unlimited. A representative set of the most common actions:

Agent actionFree / monthStarter / monthPro and up
Create task50200Unlimited
Add subtasks4 calls50 callsUnlimited
Create epic10100Unlimited
Create milestone10100Unlimited
Create sprint with tasks330Unlimited

Other write actions follow the same pattern. Comment creation is also capped, with a deliberately high ceiling even on paid plans to prevent spam. For the exact, current cap on any action on your plan, a connected agent can call the describe_capabilities tool, which reports the live limit for every tool.

Three separate limits govern AI agents. They are easy to confuse, so each is owned by its own article:

  • Action limits (this page) cap how many write actions an agent performs per month on Free and Starter.
  • Connection limits cap how many agents can be connected at once (Free 2, scaling up by plan). See Connect an external AI agent.
  • Token budget is your prepaid AI allowance that every AI feature consumes. See Understand your AI token budget.

Do these caps limit what I do by hand? No. The caps apply only to actions an AI agent performs. Creating tasks, comments, or anything else yourself in the app is never capped.

What happens when an agent hits a cap? The single action is refused with a message naming the tool and your plan. The agent can keep reading and can perform other, uncapped actions. To lift the caps entirely, upgrade to Pro or above.

Are read actions ever limited? No. Listing, getting, and searching are unlimited on every plan, including Free.

How do I see the exact limit for a specific action? A connected agent can call describe_capabilities, which returns the current monthly cap for each tool on your plan.