Getting Started with Microsoft Foundry models in VS Code

Microsoft Foundry is quickly becoming the engineering‑grade backbone for building, deploying, and operationalizing AI systems at scale. Unlike traditional “chatbot builders” or UI‑only agent platforms, Foundry is designed for real developers — people who care about versioning, reproducibility, security boundaries, and clean architecture.

At its core, Foundry gives you a unified place to define models, agents, tools, workflows, and evaluations, all backed by Azure’s enterprise‑grade infrastructure. But the real power shows up when you bring Foundry into Visual Studio Code, where everything becomes code‑first, scriptable, automatable, and production‑ready.

In this blog, we’ll walk through the setup steps to connect and use your Microsoft Foundry LLm models in VS Code. 

Prerequistes:

• Download and install VS Code - https://code.visualstudio.com
• Set up GitHub Copilot in VS Code - https://code.visualstudio.com/docs/copilot/setup
• Create a Project for Microsoft Foundry - https://learn.microsoft.com/en-us/azure/foundry/how-to/create-projects
• Deploy Microsoft Foundry Models in the Foundry portal - https://learn.microsoft.com/en-us/azure/foundry/foundry-models/how-to/deploy-foundry-models

Step 1: Install Required VS Code extensions

Install these extension on your VS Code:

• Azure Resources
• AI Toolkit for Visual Studio Code
• Microsoft Foundry

Step 2: Sign in to Azure from VS Code

1. Open the Azure resources extension in VS Code, and Sign in to Azure. 
2. Sign in using your Microsoft account (same account used for Project creation and Model deployment in Microsoft Foundry). Complete multi-factor authentication if required.
3. Choose Sign in for this app only (recommended)
   
   
4. Set the Default Microsoft Foundry Project.
5. VS Code should now be connected to your Azure account.
6. Click on the Microsoft Foundry dropdown, and verify that you can see your Foundry project(s).
   

   
7. Right Click on a project, and select Open in Foundry Toolkit Extension (a.k.a AI Toolkit)
8. In the A.I Toolkit, Expand My Resources > {yourproject} > Models section. Verify that you can see the models that you deployed on your Foundry project.
   Note: It can take a few minutes for the models to load. 
   

   

Step 3: Add your models to GitHub Copilot

1. Open the GitHub copilot chat window, and click on the Pick model button.
2. Next, click on the settings cogwheel.
   
   
3. On the 'Language models' window, Expand the Microsoft Foundry via AI Toolkit section, and search for your model name.
   

   

   
   
4. Right click on the name of model that you deployed on , click on Pin model.
5. Close the 'Language models' window.
6. Click on the Pick modelU button again, and select the model that you pinned earlier from the 'Language models' window.
   
7. You should now be able to chat and interact with you Microsoft Foundry model directly from the GitHub Copilot chat window in your VS Code.
   
   

With everything now connected, your setup elevates Foundry from a portal‑only experience to a true developer‑first workflow, letting you build, test, refine, and iterate entirely inside your editor.

From here, you’re ready to start exploring agent interactions, trigger workflows, and gradually move toward a fully code‑driven AI development process. This foundation sets you up for everything that comes next in your Foundry journey.