Attribution
OpenOrange attributes usage by instance, actor, agent, model route, request, token count, cache behavior, provider cost, and tariff context.
// account.model_usage
Model usage accounting
OpenOrange makes model calls accountable by actor, agent, route, model, request, tokens, cache behavior, provider cost, price version, customer tariff, and audit history.
Model usage accounting in OpenOrange connects every model call to the operating context that produced it: actor, agent, route, model, tokens, cache, provider price, customer tariff, approval history, and instance proof.
01
Token and prompt-cache savings stay visible per operated surface.
02
Provider costs are tied to model routes and historical price versions.
03
Customer tariffs make usage explainable instead of hand-waved.
04
Request traces connect spend to actor, agent, approval, and audit context.
Pricing history
Historical model prices and customer tariffs keep billing explainable even when provider pricing or internal commercial terms change over time.
Operational proof
Usage accounting belongs beside health, privacy, requests, approvals, applies, and audit records because spend is part of operating the AI system.
// questions
Is this a generic billing dashboard?
No. It is model usage accounting tied to private AI infrastructure, runtime context, request traces, approval history, and audit proof.
Can cache savings be explained?
Yes. The accounting model can preserve prompt-cache behavior and savings so operators can explain both spend and efficiency.