Reference Docs
FocusAI usage is metered in credits. Each plan includes a monthly allowance, and you can top up within a cycle if you run out.
Because neither of those is honest.
A "message" isn't a fixed amount of work — asking FocusAI to plan your week costs considerably more than asking it what's due tomorrow. Charging both the same would either overcharge the short question or undercharge the long one.
Raw model tokens are worse. The same token costs a different amount depending on which model handled it, so a token allowance would have meant wildly different real value depending on which model your request happened to use. Worse, when we made FocusAI dramatically more efficient, a token meter passed none of that saving on to you — the efficiency was invisible to the person paying for it.
Credits track the real cost of the work done. When FocusAI gets cheaper to run, your allowance goes further.
Free trials include Pro's allowance for the duration of the trial.
Allowances reset at the start of each billing month and don't roll over.
If you run out mid-cycle, top-up bundles are a one-time purchase on top of a paid plan:
Top-up credits are used only after your monthly allowance is exhausted, and they don't expire at the end of the month.
Settings → Usage shows credits used this month, your limit, and what's remaining — including any top-up balance. The remaining figure is the one to watch, since it already accounts for top-ups.
Settings → Models lets you pick which AI model handles your requests. Stronger models cost more credits per request; if you're planning a complex quarter, that's worth it, and for a quick lookup it isn't.
Members of an organization get an allowance according to their seat. Admin seats are provisioned at the Plus level and member seats at the Pro level.