Usage And Limits
Customer Line Items
Customer-facing reports roll detailed platform usage into readable line items:
- AI inference
- Generation jobs
- Runtime compute
- Storage
- Bandwidth
- Subscription
Internal provider names, platform overhead, and reconciliation details are not shown in customer invoices.
Usage Bars
Usage bars show:
- current usage
- reserved in-flight usage where applicable
- included limit
- warning thresholds
- grace status when the plan supports grace
Limits
Some actions reserve budget before they start. Examples include AI calls and generation jobs. If the operation fails, the reservation is released. If the operation costs more than estimated, Gessa tries to reserve the additional amount before continuing.
High-volume usage such as runtime minutes, player minutes, bandwidth, and storage is metered asynchronously and appears in rollups.
Guest Usage
Guest usage is attributed to the workspace that owns the public deployment. Guest sessions have plan-specific limits for session duration, concurrency, bandwidth, and expected cost.