Multi-tenant cloud
The default. Shared gateway and router, dedicated project isolation, nine regions. Nothing to operate.
Five components sit between your application and the models. Each one is independently scaled, independently deployable, and degrades to something useful rather than to an error.
A provider outage is a routine event, not an incident. The router keeps live health probes against every model in your pool and removes a degraded endpoint from rotation before your error rate moves.
A single request that hits a degraded provider, told in the order it happens.
frontier-lg — cheapest model clearing the quality floor.reason-pro, the next entry in the chain.route.attempts: 2. Your code sees one successful call.Pin inference and storage to a region to satisfy data-residency requirements. Private deployments add single-tenant regions on request.
| Region | Location | Inference | Trace storage | Status |
|---|---|---|---|---|
| us-east | N. Virginia | Active | Active | Operational |
| us-west | Oregon | Active | Active | Operational |
| eu-central | Frankfurt | Active | Active | Operational |
| eu-west | Dublin | Active | Active | Operational |
| uk-south | London | Active | Active | Operational |
| ap-northeast | Tokyo | Active | Active | Operational |
| ap-southeast | Singapore | Active | Active | Operational |
| ap-south | Mumbai | Active | Regional | Operational |
| sa-east | São Paulo | Active | Regional | Operational |
The default. Shared gateway and router, dedicated project isolation, nine regions. Nothing to operate.
Your own gateway, router, and trace store in a region we operate. Same API, isolated infrastructure and upgrade schedule.
Run the gateway and router inside your own network. Model traffic egresses only to the providers you allow.
The engineering team publishes design notes as systems change.