Text features spend tokens. A short answer over a few retrieved passages is a different bill from a long agent that keeps calling tools. Voice spends phone minutes as well as tokens, and a minute of a call is often the number that surprises people.
We estimate that cost before the sprint, from the traffic you already have: requests a tenant makes in a busy week, how long a typical context is, and, for voice, how long a call tends to last. After launch, each tenant has a cap. If usage runs away, the feature stops before the invoice does. The provider bills your account.
Bring your traffic to a call and we will walk through the estimate with you. There is no dollar estimator on this page.
The feature this is for is described on the AI features page.
The running cost depends on your traffic and the feature. We estimate it in the scope. We do not publish a starting price.
You do. The provider bills your account. We put a cap on each tenant so a busy week does not become a surprise.
