A recap of the most frequently asked questions about the Model Context Protocol (MCP) from my time as an expert at the MCP booth during Microsoft Build 2025. Covers what MCP is, how it works, authentication, differences from OpenAPI, agent-to-agent communication, and practical advice for developers.

In this blog post, I’ll share how I integrated the Model Context Protocol (MCP) into a real SaaS product, VocalCat, by reusing an existing API. This guide will highlight the interesting bits, challenges, and tips I discovered along the way. Buckle up! Why MCP? MCP is an open standard that simplifies connecting AI models to external tools and data systems. In plain terms, MCP is a way to give “hands” to copilots—a protocol that lets AI models not just answer questions, but actually take actions, even in the real world. Imagine connecting your copilot to your IoT devices, so it can control things in your home or office. It’s a bit dangerous, but incredibly exciting: suddenly, copilots can do more than just talk—they can act.