AI assistants are only as powerful as the tools they can access. Model Context Protocol (MCP) servers allow you to expose your own tools, data, and workflows to AI-but how do you actually build one?

In this session, we’ll create a working MCP server from scratch using .NET and the MCP SDK. Starting with a simple project, we’ll explore the core building blocks of the Model Context Protocol; resources, prompts, and tools, and see how they enable AI assistants to access custom capabilities and data.

Then we’ll go further than a local demo. Using a practical sample inspired by the Azure remote MCP functions approach, we’ll deploy our server to the cloud and connect it to an AI client, turning a simple .NET service into something AI agents can use.

Along the way, we’ll look at common pitfalls, debugging techniques, and how MCP servers fit into modern AI workflows.