MCP_POC_With_ChatTool

# MCP_POC - Simple Azure OpenAI tools demo (.NET 8) This sample shows a minimal, console-based "MCP-like" experience using Azure OpenAI function tools in .NET 8. What it does: - Starts a console app - Defines two tools: Temperature of city and add numbers - Sends your input to Azure OpenAI and executes tools when requested ## Prerequisites - .NET 8 SDK installed - An Azure OpenAI resource with: - Endpoint URL (e.g., https://<your-resource-name>.cognitiveservices.azure.com/) - API key - A chat model deployment name (e.g., model-router, gpt-4o-mini) ## 1) Get the code - Open this folder in your terminal/VS Code/Visual Studio. ## 2) Install packages (commands) Run these in the project folder (where MCP_POC.csproj is): - Required - dotnet add package Azure.AI.OpenAI --version 2.1.0 - Optional (only if you switch to DefaultAzureCredential later) - dotnet add package Azure.Identity --version 1.14.2 ## 3) Configure Azure OpenAI settings In ChatToolForMCP.cs, update the three values: - endpoint: set to your Azure OpenAI endpoint URL - apiKey: set to your Azure OpenAI API key - deploymentName: set to your chat model deployment name (for tools support, use a modern chat-capable model) Example (replace placeholders): - endpoint = "https://YOUR-RESOURCE.cognitiveservices.azure.com/" - apiKey = "YOUR-API-KEY" - deploymentName = "YOUR-DEPLOYMENT-NAME" Tip: Do not commit real keys to source control. Prefer environment variables or user-secrets in real apps. ## 4) Build - dotnet build ## 5) Run - dotnet run You should see prompts like: - "MCP Server initialized with tools:" - Enter your text, for example: - Echo something: echo "hello world" - Add numbers: "add 5 and 7" ## Troubleshooting - 401 Unauthorized: check API key - 404/BadRequest: check deployment name and that the model supports tools - Timeouts/Network: check your firewall/proxy and endpoint URL ## Notes - Target framework: .NET 8 - Packages kept minimal for clarity - Keep secrets out of source code in real projects