Your monthly allowance, the caps that protect it, what happens when you hit one, and how to watch your usage.

One allowance, resets every month

Flat tiers (Pro) start each billing cycle with a fresh monthly allowance of spendable credit at raw provider cost (0% token markup), covering Budget, Mid, and Premium models — Ultra-tier models draw your prepaid balance instead, not the allowance. The allowance resets on your subscription's renewal date each cycle, not the calendar 1st, and any unused portion is forfeited — it does not carry over to the next cycle. PAYG has no fixed allowance; it is prepaid at raw provider cost (0% markup) with no monthly fee.

Cycle-anchored reset. The allowance resets on your subscription's renewal date each month, not the calendar 1st — even on annual billing, it still resets monthly on that cycle anchor. Unused allowance does not carry over.

Caps per model tier class

On top of the overall allowance, each customer has weekly and monthly token caps per model tier class (budget / mid / premium / ultra). Lower tiers can't reach Premium or Ultra classes, and even Pro can't exceed its class caps. Hitting a cap returns 402 token_quota_exceeded with the class and window that tripped.

402 response

HTTP/1.1 402 Payment Required
content-type: application/json

{
  "detail": {
    "error": "token_quota_exceeded",
    "model_tier_class": "premium",
    "window": "monthly"
  }
}

The model_tier_class field tells you which model class hit the cap; window is either weekly or monthly. Switch to a lower-cost model class or raise the cap with a token pack.

What happens when the allowance runs out

When a flat tier's allowance is exhausted, the outcome depends on a per-account overage toggle:

Overage enabled: further usage draws down your prepaid balance instead of being blocked.
Overage disabled: the request is blocked with 402 until the next reset or a top-up.
PAYG never blocks on quota — it has no fixed allowance and is always prepaid at raw provider cost (0% markup).

RPM / TPM are separate from quota

Rate limits are not the same thing as your token quota. RPM (requests/min) and TPM (tokens/min) are anti-flood burst guards that return 429 rate_limit_exceeded — a temporary throttle, not a spend cap. Your monthly token quota returns 402 token_quota_exceeded instead. Burst credits (token packs) can temporarily raise a class's token cap from prepaid balance.

See Rate Limits for per-tier RPM/TPM/inflight numbers, the response headers, and backoff patterns.

Watch weighted tokens live

The dashboard Usage page shows weighted tokens used vs. allowance and the reset date. Programmatically, call the subscription quota endpoint with your API key.

curl

curl https://api.anoman.io/anoman/v1/subscription/quota \
  -H "Authorization: Bearer anm-sk-..."

Enterprise and PAYG return a limit of 0, meaning no fixed monthly cap applies.

Watch your quota

The Usage page shows exactly how many weighted tokens you have left and when they reset.

Quota & overage