AI Gateway

AI Gateway routes app AI requests through server-side provider configuration and exposes usage controls.

Before you start

Obtain AI Gateway permission, an app-scoped API key and either the platform AI access available to your account or your own provider configuration. Feature availability and model capabilities depend on the provider and plan.

Step by step: configure

Open the app's AI Gateway. Resolve App API Key Missing through the private API Keys page if needed.
Inspect whether the app inherits profile AI or uses Override profile AI for this app.
Choose the available Chat, Image, Vision and Audio providers/models for the capabilities you need.
Enter provider secrets only in the private server-side configuration fields. Never embed them in a mobile app or a public prompt.
Review any custom gateway URL or system-prompt override. Use only an endpoint you trust with the data it will receive.
Under Usage Limits & Controls, set the limit type and monthly per-user limit to suit your intended usage.
Select Save Changes, then Test connection. A test may call the selected provider and consume usage.

Step by step: integrate and monitor

Use the corresponding SDK guide or private API Explorer with YOUR_API_KEY replaced only in your own environment.
Start with a small, non-sensitive request for the chosen capability.
Inspect Request Logs, filtering Success or Errors.
Review traffic and Estimated spend by model. Estimates are not provider invoices.
Confirm the app receives and handles both successful and failed responses.

Confirm it worked

A test request succeeds with the intended provider/model and the usage log matches it. Saving configuration alone does not prove every modality is supported.

Troubleshooting

Check app key scope, provider/model access, limits, endpoint configuration and error messages. Do not assume “AI included” means unlimited usage or that enabling the gateway installs AI features into your app.

Related: API keys, API Explorer, Billing.