Every inference request, routed right.

Proximarun dispatches each AI inference call to the node that fits your latency and cost target. Under 12ms of routing overhead. No manual rules.