---
title: "AI inference"
url: "/guides/inference/"
kind: "hub"
summary: "Route to self-hosted model servers on Kubernetes with Gateway API, InferencePool, and an external endpoint picker."
---

> Arion site index: [/llms.txt](/llms.txt)

# AI inference

Route to self-hosted model servers on Kubernetes with Gateway API, InferencePool, and an external endpoint picker.

Configure Kubernetes Gateway API resources so Arion routes inference requests to ready model server Pods selected by an external endpoint picker.