Limits and cost controls

Three layers keep usage and cost under control: plan capacity, spend limits you set yourself, and fixed safety ceilings. Nothing is ever deleted when a limit is hit; work pauses and resumes when the limit is raised or the month rolls over.

Plan capacity

Your plan sets how many workspaces, agents, and reports the billing account can hold, pooled across its workspaces (see Plans & AI credit).

• At the limit, creating another one is refused with a clear message. Free a slot (delete or deactivate something) or upgrade the plan.
• Report runs are a pooled monthly allowance. Once it is exhausted, further scheduled runs are recorded in the run history as limit reached and skipped. Schedules resume automatically with the new month.

Spend limits

Spend limits are monthly caps in euros on billable AI work (chat and agent runs). They are checked before each run against the same live figures the usage dashboards show.

• Workspace limit. Caps a workspace's AI spend per month. New workspaces start with a 100 € limit. Edit or remove it under Settings → Cost Management → Limits (workspace admins only). Removing it means the workspace is limited only by the credit balance.
• Per-agent limit. Every agent has its own monthly cap, 10 € by default, on its Usage tab. This makes a runaway automation a bounded problem: one agent hitting its cap stops that agent, not the workspace.

When a limit is reached, agent runs are recorded as limit reached and skipped, and chat responds with a clear notice instead of starting work. Raise or remove the limit to continue immediately, or wait for the month to roll over. Setting a limit to 0 is a deliberate pause.

Fixed ceilings

A few per-workspace safety ceilings are the same on every plan: 30 members, 50 channels, 10 recipients per email channel, 200 schedules, and 5,000 report deliveries plus 5,000 agent deliveries per month. They exist to catch configuration mistakes, not to be reached in normal use.