Tetrate Adds Token Brokering for AI Cost Management in Agent Router Enterprise

Tetrate has added a token brokering capability to its Agent Router Enterprise platform, announced in a press release. The new feature allows AI and platform teams to manage token spending across distributed models using a command line interface, software development kit, and a set of APIs.

The system introduces a distributed inference control plane that connects developers and agents to the models they use. Administrators can set budgets, define approved models and regions, and establish fallback rules. Agent Router Enterprise enforces these rules automatically on every request, redirecting workloads to cost-effective alternatives when budgets are reached.

Built on the Envoy AI Gateway project, the updated system enables governance and budget control across frontier, private, and edge models. It is designed to integrate into existing workflows, separating administrative policy management from developer operations.

The new token brokering functionality is generally available to all Agent Router Enterprise subscribers at no additional cost.