- Demo images: limits, if any, are set for the pilot and communicated directly.
- Licensed production images: throughput is bound by your own GPU, not by MWire. Concurrency limits are a config value in your runbook, not a contract term.
- 429 responses only occur if you configure a limit yourself, or during a time-boxed demo/pilot.
Ops
Rate limits
No public default. Set per license or pilot agreement.
There is no fixed public rate limit. This is an on-prem, licensed API, not a shared multi-tenant service.
