Agentic AI is only as powerful as the tools it can safely call, and Docker’s new MCP Gateway and Toolkit make that integration dramatically easier and more secure.

In this session, we'll explore how core concepts of MCP, then show how Docker’s MCP Gateway provides a unified, secure endpoint for your agents while the MCP Toolkit simplifies discovery, configuration, and lifecycle management of MCP servers inside containerized environments. Using live demos, we’ll wire an LLM-based agent to real-world tools (APIs, search, and code repositories), run everything locally with Docker, and then discuss how to evolve the same setup toward team and production use.

Target audience: developers and architects who are already experimenting with LLMs or AI assistants and want a practical path to building maintainable, debuggable, and secure agentic applications with Docker.