Usage limits
Steinkauz AI Usage limits let you set cost limits and token limits for each member or dedicated API key on each provider instance. They are optional policy ceilings, not included inference, and not a commercial seat meter.
Where to configure
- Usage limits: organization owners and admins set optional caps per provider.
- API access: members create keys whose requests count towards their personal usage limits. Only owners and admins can select Dedicated usage attribution or manage separate limits for a key.
Private Deployment usage limits are available without a Cloud subscription. You can open Usage limits and set optional money or token caps; they reset on the UTC calendar month.
Usage limits are optional in both Cloud and private deployments. Leave a cost or token limit empty to operate without that cap. Configured limits are always enforced, and zero means exhausted. Usage is recorded whether or not you configure limits.
Optional BYOK caps
On each provider tab, admins set optional:
- Members: each member’s cost limits and/or token limits
- API keys: separate cost limits and/or token limits for keys with Dedicated usage attribution
Empty means uncapped; 0 means immediately exhausted. Edits that would set a cap below that subject’s spend this period are rejected.
Usage attribution
| Usage attribution | Which limits apply |
|---|---|
| Inherit (default) | The creator’s personal usage limits |
| Dedicated | Separate cost limits and token limits for the key (empty = uncapped) |
Owners and admins can switch modes after creation. Spend history stays on the member or key that recorded it. Switching a dedicated key to inherit mode, or revoking it, removes its key-specific usage limits.
Exhaustion
When any configured unit (money, input tokens, or output tokens) has Remaining ≤ 0, further inference for that subject on that provider is blocked. The Platform API returns 402 insufficient_quota on the next attempt after an overshooting call (that call still completes with HTTP 200).
A member or dedicated key without configured usage limits on a provider has no optional cost or token cap there. Owners and admins can configure these caps on Usage limits.
A customer-owned Vercel AI Gateway connection still needs your Gateway API key.
Other providers (BYOK)
Each member can have optional Cost ($), Tokens In, and Tokens Out caps. Token caps show compact counts (K/M/B) until you focus the field. Period spend shows used/cap progress when a track is set.
If the provider reports tokens but not cost (or the reverse), a warning appears once for that provider: the corresponding cost or token limit cannot be enforced until the provider reports that unit. Saving never blocks on that warning.
Use the save bar to apply all unsaved usage limit changes for the active provider.
Configuration Audit
Cap edits appear in Change history. The Usage limits page links to filtered history. See Configuration Audit.