Billing has three moving parts: the plan, the seats, and the metering. Plans are per store — each connected store carries its own plan, with a seat count for the operators who work in it. For current plan tiers and prices, see the pricing page; this article covers how the mechanics work regardless of tier.
What an AI-credit is
Usage is metered in AI-credits. Each model or API call an agent makes consumes credits, and every plan includes a monthly credit allotment with soft overage — you're not cut off mid-task, and the meter is visible so 100% is never a surprise. Because every call is metered, expected usage tracks how hard your agents actually work, not a per-seat abstraction. See the pricing page for what each plan includes.
LLM spend is separate and capped
Raw model spend is additionally governed by the per-store monthly LLM budget cap, not by your plan tier. Once a store hits its cap, further LLM work is skipped rather than spending through the ceiling. Credits meter what you use; the budget cap bounds what agents can spend. The two limits are independent on purpose.
Invoices, VAT, downgrades
- Invoices live in the billing tab, with VAT handled per your billing country — enter a valid VAT number and reverse-charge applies where applicable.
- Upgrades take effect immediately, prorated.
- Downgrades take effect at the end of the current billing period — you keep what you paid for until the period closes, and nothing is clawed back mid-cycle.
If your usage pattern doesn't fit the metering — very high credit consumption, agencies running many stores — talk to us before the invoice does the talking.