What’s new in Unity AI Gateway: service policies, guardrails, observability, and cost controls for AI agents and MCPs

Summary

Unity AI Gateway is expanding runtime governance for production AI agents, model calls, and MCP tool interactions as teams face rising costs, unclear behavior, and limited control. The Beta adds LLM-based guardrails, token-level cost attribution with per-user alerts and hard budget limits, payload logging, and MCP service policies. Guardrails use a model and prompt to evaluate inputs, outputs, or both in real time, while inference tables and Unity Catalog system tables centralize governed records of usage and interactions. For MCPs, administrators can define SQL policies as Unity Catalog functions using agent identity, user context, and request parameters to constrain tool access and sensitive actions. The capabilities are available in Beta, with payload logging and service policies offered through gated enrollment, and are intended to improve observability, compliance, and cost control.

Takeaways

1. 1

   LLM-based guardrails can evaluate model inputs, outputs, or both in real time using customizable model-and-prompt policies for safety, PII protection, and business rules.

2. 2

   Cost controls provide token-level attribution across requests, users, and endpoints, plus per-user alerts and hard budget limits across multiple models and providers.

3. 3

   MCP service policies are defined as Unity Catalog functions and can use agent identity, user context, and request parameters to restrict tools or require approval for sensitive actions.