Public Inference

Platform administrators manage two kinds of public models in Platform Management → Public Inference: in-cluster local deployments, and cloud API channels that forward to external vendors.

Full procedures (tabs, channel fields, model mapping, connection test) are in the Industry AI Model Platform docs:

• Manage Public Inference
• Public Inference Overview
• Use Public Inference

The user-facing entry is Model Inference → Public Inference. Cloud requests go through /aigateway on the console. The request model is the platform model ID registered on the channel.