> ## Documentation Index
> Fetch the complete documentation index at: https://agumbe.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Go to Production

Before sending production traffic through Agumbe AI Gateway, make sure your integration is secure, observable, and easy to operate.

Use this checklist to move from a working request to a production-ready setup.

1. **Call the gateway from your backend**

* Production traffic should go through your backend, worker, or service layer.
* Do not expose Gateway API keys in browser or mobile clients. Your client application should call your own backend, and your backend should call Agumbe AI Gateway.

**Recommended flow:**

Client application → Your backend → Agumbe AI Gateway → Model provider

2. **Use the right API key scope**

* Choose the API key scope based on how your application uses guardrails.
* Use an app-scoped key when one workload should always use the same app policy. This is a good fit for production apps, staging apps, sensitive workflows, or services with a clear single purpose.