Microsoft Azure

​

Overview

• Azure AI Foundry — Tavily is featured in the Tools Catalog of Azure AI Foundry, making it easy for users to add Tavily to their AI agent workflows and leverage its capabilities for real-time information retrieval and grounding.
• Azure MCP Center — Tavily Remote MCP is listed on Azure’s MCP Center (part of Azure API Center), enabling developers to discover and connect to Tavily directly.

​

Prerequisites

• Microsoft Azure account with active subscription for signing in to Azure AI Foundry
• Tavily API Key for connecting the Tavily Remote MCP to Azure AI Foundry

​

Setup

​

Azure AI Foundry

1

Visit Azure AI Foundry

Go to Azure AI Foundry and sign in with your Microsoft Azure account.

2

Toggle on the New Foundry

On the top bar, toggle on the New Foundry to switch from Microsoft Foundry (classic) to Microsoft Foundry (New).

3

Select or create a new project

A pop-up will appear to select an existing project or create a new one. Select the project you want to work on or create a new project to get started.

4

Enter Project Details

Fill in the required project details such as Foundry resource, subscription, region, and resource group to set up your project environment to create a new project in Microsoft Foundry(new).

5

Create an agent

Once your project is set up, click on the Start Building button to create a new agent within your project.

6

Add Tavily MCP to your agent workflow

In the playground interface, navigate to the Tools section and search for Tavily MCP in the catalog. Enter your Tavily API Key to authenticate and connect your agent to the Tavily Remote MCP. Once connected, Tavily MCP will be available as a tool in your agent’s workflow for real-time web search and RAG capabilities.

7

Start using Tavily MCP in your agent workflow

With Tavily MCP added to your agent’s tools, you can now use it within your agent’s workflow to enhance its capabilities with real-time information retrieval and grounding for more effective AI agent performance.

​

Azure MCP Center

1

Visit Azure MCP Center

Go to the Azure MCP Center and search for Tavily MCP to find the listing.

2

Install in VS Code

Click the Install button on the Azure MCP Center listing for Tavily MCP to add it to your VS Code.

3

Authenticate with Tavily Account

VS Code will prompt you to authenticate with your Tavily account to connect to the Remote MCP. Follow the authentication flow to grant access.

4

Start using Tavily MCP with Copilot

Use Tavily MCP within VS Code’s Copilot to enhance your AI agent development with real-time web search and RAG capabilities.

​

Resources

• Build a workflow in Microsoft Foundry
• Use MCP servers in VS Code
• Tavily MCP Documentation