Every AI platform has limits. The difference is what happens when you reach one.

The common failure mode: a request quietly dies, an error code appears, and you are left guessing whether the model is down, your account is broken, or the weather is bad.

Our rule: the refusal explains itself

When a ToRun request is declined for budget reasons, the message tells you three things:

• What limit you reached — your included allowance for the current session window, your wallet balance, or a plan ceiling.
• When it resets — session allowances refill on a rolling window, not at an arbitrary midnight. The message shows the actual time.
• What your options are — wait for the reset, top up your wallet, or move to a plan that fits your usage. Each option is a link, not a scavenger hunt.

Why a rolling window

Daily quotas punish time zones and reward quota-camping at midnight. A rolling session window means your allowance recovers continuously as older usage ages out — fair at 9am and at 11pm, on every continent.

The quiet part

Budget refusals are not a sales funnel. The upgrade suggestion appears when it is genuinely the answer to "how do I keep working right now" — and the free option (wait for the reset) is always stated first.