After successful installation of the AI Guardrails Service certificate, follow the steps below to Configure AI Guardrails Service Host.
-
In the AI Guardrails Service page, select Configure AI Guardrails Service Host.
-
If no host is configured before, the Configure AI Guardrails Service Host page displays No host configuration found message under the prompt Current Host Configuration.
-
Use the up arrow key or down arrow key to navigate between the parameters.
-
Enter values for the following parameters.
- Enter new Host URL (required parameter): your LLM backend URL.
For Example: https://<llm-guardrails-endpoint>/invocations).
NOTE: You must append the specific inference path suffix/invocationsto the end of your URL. - Enter new JWT URL (optional parameter): OAuth2 token endpoint. Example: https://<auth-provider>/oauth2/token
- Enter new client ID (optional parameter): OAuth2 client ID from your OAuth2 provider, or service account email for TPU deployments.
- Enter new client secret (optional parameter): OAuth2 client secret from your OAuth2 provider, or RSA private key for GCP TPU deployments.
- Enter new scope (optional parameter): OAuth2 scope for token requests, or JWT audience for TPU deployments.
- Enter new Host URL (required parameter): your LLM backend URL.
-
Press Enter to save the AI Guardrails service host.
The system validates the configuration and saves it. If validation fails, press R to retry the configuration or press Esc to return to the previous menu. On successful installation, the Configure AI Guardrails Service Host page displays a completion message and provides the host details.