AI Agents with MCP: A Practical Guide to Building Clients, Servers, and End-to-End Agents

Summary

Kyle Stratis’s 332-page O’Reilly book is a practical guide to using the Model Context Protocol (MCP) to build AI agents and the applications that connect them to tools. It explains where MCP came from and how clients, servers, and transports fit together, then guides readers through implementing a client and a server in Python with support for the full specification. The server projects provide tools, prompts, and resources, while later chapters cover advanced client and server use, utilities, and integrating client capabilities. The book also addresses testing, securing, and sharing MCP servers, along with the details of connecting clients and servers through MCP transports. Its final chapter discusses extensions, tools for building with MCP, and contributing to the protocol. The ebook is available now, while the print edition is available for pre-order through Amazon, O’Reilly, Barnes & Noble, and Google Books. The author, Kyle Stratis, is a software engineer and consultant whose work spans AI engineering, computer vision, data engineering, and machine-learning engineering.