> For the complete documentation index, see [llms.txt](https://docs.katara.ai/about/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.katara.ai/about/platform/ai-gateway.md).

# AI Gateway

The AI Gateway keeps model traffic inside policy. It is the component that routes AI requests through Katara so each request can carry the right organization boundary, access policy, budget, and audit trail.

## What it governs

The AI Gateway is documented as the place where model and tool traffic can be authenticated, evaluated, routed, metered, and recorded.

It helps teams answer questions such as:

* Who or what made the request?
* Which organization, user, or application grant was in scope?
* Which model, endpoint family, server, or tool was involved?
* What was the outcome, latency, usage, spend, or budget status?

## How to think about it

Katara turns each AI request into a governed transaction. Applications call Katara first; Katara checks the relevant boundary and policy before traffic continues to approved upstream systems.

Provider setup, SDK examples, and configuration details are intentionally not covered here.