Radiant is a gateway for AI applications. It supports companies building AI applications that use multiple models and operate across environments (cloud, on-prem, hybrid-cloud), with multi-model governance and management. Radiant allows easy switching between any number of locally-run and external models while not changing code or system architecture. Built from the ground-up to support air-gapped AI, Radiant runs entirely inside a customer's network and manages AI workloads across environments.