Every call the assistant makes is logged: which provider and model, what it was for, how long it took, roughly how many tokens, and an estimated cost. No prompts and no replies are stored in that log, only the shape of the call.
- On the Dashboard: calls, estimated cost and errors for the date range on screen, the split by job (answering chats, opening lines, screening messages), the average reply time, and this month's spend against your budget.
- Under the AI Assistant tab: the last ten calls with any error message word for word.
- Monthly soft budget. Set an amount in USD. When this month's estimate reaches it, the assistant stops answering and screening until next month or until you raise it. Chat itself carries on, person to person. Set 0 for no limit.
The cost is an estimate worked out from the provider's published rates and the length of each message. Your provider's own billing page is the real figure. A Gemini free key normally shows a cost here while costing you nothing, because the estimate does not know you are on the free tier.
Only an administrator of this site ever sees these figures. Visitors and customers never do.