Fleet + Compute: GPU nodes meet the cloud API

Fleet is our control plane for distributed agent workers — register nodes, dispatch jobs, and monitor runs from a single dashboard. Compute adds private GPU capacity for inference and training workloads that should not leave your environment.

Both products connect to the same Simon API surface your application already uses. That means a chat app, CAT workflow, or Tiger Mode agent can route heavy work to dedicated hardware without a second integration.

Early customers use Fleet for long-running automations and Compute for fine-tunes or high-volume inference where cloud egress and latency matter.