Foundry Local for Microsoft Sovereign Private Clouds

Deploy and run AI inference workloads on Azure Local with Arc-enabled Kubernetes. Use OpenAI-compatible APIs with your AI models while keeping data on-premises.

About Foundry Local for Sovereign Private Clouds

Overview

• What is Foundry Local on Azure Local?
• Known issues

Key concepts

Concept

• Models
• Model catalog and sourcing
• Bring your own models
• Inference operator and model lifecycle
• Model caching and StoreModel lifecycle
• Inference runtimes
• Authentication and authorization
• Automatic GPU inference tuning

Get started

Quickstart

• Deploy your first model and run inference

Deploy

• Configure TLS
• Deploy Foundry Local as an Azure Arc extension
• Deploy a catalog model
• Deploy a bring-your-own model

Secure

How-To Guide

• Configure Entra ID authentication

Run inference

How-To Guide

• Run inference

Disconnected operations

Overview

• Overview

How-To Guide

• Prepare deployment
• Deploy Foundry Local as an Azure Arc extension
• Deploy your first model

Reference

Reference

• vLLM model reference
• Inference API reference
• Inference API endpoints and payload reference
• ModelDeployment and operator configuration reference
• Glossary