name	description	languages	products	page_type	urlFragment
.NET OpenAI MCP Agent	This is an MCP agent app written in .NET, using OpenAI, with a remote MCP server written in TypeScript.	csharp bicep azdeveloper	azure-openai azure-container-apps azure	sample	openai-mcp-agent-dotnet

.NET OpenAI MCP Agent

This is an MCP agent app written in .NET, using Azure OpenAI, with a remote MCP server written in TypeScript.

Features

This app provides features like:

• The MCP host + MCP client app is written in .NET Blazor.
• The MCP client app connects to a to-do MCP server written in TypeScript.
• Both MCP client and server apps are running on Azure Container Apps (ACA).
• The MCP client app is secured by the built-in auth of ACA.
• The MCP server app is only accessible from the MCP client app.

Prerequisites

• .NET 9 SDK
• Visual Studio Code + C# Dev Kit
• node.js LTS
• Docker Desktop or Podman Desktop
• Azure Subscription

Getting Started

You can now use GitHub Codespaces to run this sample app (takes several minutes to open it)! 👉 .

Get Azure AI Foundry or GitHub Models

• To run this app, you should have either Azure AI Foundry instance or GitHub Models.
• If you use Azure AI Foundry, make sure you have the GPT-5-mini models deployed deployed.
• As a default, the deployed model name is gpt-5-mini.

Get AI Agent App

1. Create a directory for the app.

   # zsh/bash
   mkdir -p openai-mcp-agent-dotnet

   # PowerShell
   New-Item -ItemType Directory -Path openai-mcp-agent-dotnet -Force

2. Initialize azd.

   cd openai-mcp-agent-dotnet
   azd init -t openai-mcp-agent-dotnet

   NOTE: You'll be asked to enter an environment name, which will be the name of your Azure Resource Group. For example, the environment name might be openai-mcp-agent.

3. Make sure that your deployed model name is gpt-5-mini. If your deployed model is different, update src/McpTodo.ClientApp/appsettings.json.

   {
     "OpenAI": {
       // Make sure this is the right deployment name.
       "DeploymentName": "gpt-5-mini"
     }
   }

Get MCP Server App

1. Clone the MCP server.

   git clone https://github.com/Azure-Samples/mcp-container-ts.git ./src/McpTodo.ServerApp

2. Set JWT token.

   # zsh/bash
   ./scripts/set-jwttoken.sh

   # PowerShell
   ./scripts/Set-JwtToken.ps1

Run on Azure

1. Check that you have the necessary permissions:

   • Your Azure account must have the Microsoft.Authorization/roleAssignments/write permission, such as Role Based Access Control Administrator, User Access Administrator, or Owner at the subscription level.
   • Your Azure account must also have the Microsoft.Resources/deployments/write permission at the subscription level.

2. Login to Azure.

   azd auth login

3. Add JWT token to azd environment.

   # zsh/bash
   env_dir=".azure/$(azd env get-value AZURE_ENV_NAME)"
   mkdir -p "$env_dir"
   cat ./src/McpTodo.ServerApp/.env >> "$env_dir/.env"

   # PowerShell
   $dotenv = Get-Content ./src/McpTodo.ServerApp/.env
   $dotenv | Add-Content -Path ./.azure/$(azd env get-value AZURE_ENV_NAME)/.env -Encoding utf8 -Force

4. Deploy apps to Azure.

   azd up

   NOTE:

   1. By default, the MCP client app is protected by the ACA built-in auth feature. You can turn off this feature before running azd up by setting:

      azd env set USE_LOGIN false

   2. During the deployment, you will be asked to enter the Azure Subscription, location and OpenAI connection string.

   3. The connection string should be in the format of Endpoint={{AZURE_OPENAI_ENDPOINT}};Key={{AZURE_OPENAI_API_KEY}}. The Azure OpenAI API endpoint should look like https://<location>.api.cognitive.microsoft.com/.

   4. You can add GitHub PAT in the same format above to use GitHub Models like Endpoint=https://models.inference.ai.azure.com;Key={{GITHUB_PAT}}.

5. In the terminal, get the client app URL deployed. It might look like:

   https://mcptodo-clientapp.{{some-random-string}}.{{location}}.azurecontainerapps.io/

6. Navigate to the client app URL, log-in to the app and enter prompts like:

   Give me list of to do.
   Set "meeting at 1pm".
   Give me list of to do.
   Mark #1 as completed.
   Delete #1 from the to-do list.
   

   NOTE: You might not be asked to login, if you've set the USE_LOGIN value to false.

7. Clean up all the resources deployed.

   azd down --force --prune

Run locally

1. Make sure you're in the openai-mcp-agent-dotnet directory.

2. Add Azure OpenAI API Key.

   dotnet user-secrets --project ./src/McpTodo.ClientApp set ConnectionStrings:openai "Endpoint={{AZURE_OPENAI_ENDPOINT}};Key={{AZURE_OPENAI_API_KEY}}"

   NOTE: You can add GitHub PAT in the same format above to use GitHub Models like Endpoint=https://models.inference.ai.azure.com;Key={{GITHUB_PAT}}.

3. Run the MCP server app.

   cd ./src/McpTodo.ServerApp
   npm run dev

4. Run the client app in another terminal.

   dotnet watch run --project ./src/McpTodo.ClientApp

5. Navigate to https://localhost:7256 or http://localhost:5011 and enter prompts like:

   Give me list of to do.
   Set "meeting at 1pm".
   Give me list of to do.
   Mark #1 as completed.
   Delete #1 from the to-do list.
   

6. Type CTRL+C to stop the agent app.

Run in local containers

1. Make sure that you're running either Docker Desktop or Podman Desktop on your local machine.

2. Make sure you're in the openai-mcp-agent-dotnet directory.

3. Export user secrets to .env.

   # zsh/bash
   dotnet user-secrets list --project src/McpTodo.ClientApp \
       | sed 's/ConnectionStrings:openai/ConnectionStrings__openai/' \
       | sed 's/McpServers:JWT:Token/McpServers__JWT__Token/' > .env

   # PowerShell
   $($(dotnet user-secrets list --project src/McpTodo.ClientApp) `
       -replace "ConnectionStrings:openai", "ConnectionStrings__openai") `
       -replace "McpServers:JWT:Token", "McpServers__JWT__Token" | Out-File ".env" -Force

4. Run both apps in containers.

   docker compose up --build

5. Navigate to http://localhost:8080 and enter prompts like:

   Give me list of to do.
   Set "meeting at 1pm".
   Give me list of to do.
   Mark #1 as completed.
   Delete #1 from the to-do list.
   

6. Type CTRL+C to stop all containers.

7. Delete all running containers.

   docker compose down

Resources

• .NET AI Template
• Model Context Protocol (MCP) C# SDK
• MCP .NET samples
• MCP Todo app in TypeScript