The future of AI integration is here. Join Kent C. Dodds' remote MCP workshop cohort and master the Model Context Protocol for next-gen AI apps.
The most exciting thing in software right now? MCP.
Instead of clicking through pages and menus and forms, users are beginning to engage with their favorite applications through natural language, voice, and AI agents.
But for AI to actually do anything useful, it needs more than just a chat box. It needs a structured way to understand and interact with your app's capabilities and your users’ data.
Chat interface where Kent asks for a popular blog post recommendation; the assistant replies it can’t know what he’s read but offers to search popular or most-discussed posts instead.
That's exactly what MCP provides: a standardized interface for building intelligent software. MCP defines how apps expose tools, resources, and context in a way that large language models (LLMs) and other agents can reliably call and consume.
MCP Fundamentals - Core architecture, principles, and building blocks
Advanced MCP Features - Real-world applications and advanced techniques
Authorization - Secure, authenticated AI interaction across networks with OAuth 2.1
User Interface (MCP-UI) - Rich interactive experiences beyond text
You'll learn:
How to create servers that leverage MCP's most useful features
The vocabulary to navigate the new MCP ecosystem
How to add UI for the best user experience
Authentication and user-specific data access
The latest features and techniques for building production-ready servers
And a lot more.
Workshop 1: MCP Fundamentals
Days 1-2: Setup, Tools, Resources & Prompts
Master the core building blocks of the Model Context Protocol. You'll start by setting up your development environment and immediately dive into building functional tools that AI assistants can discover and use. Learn how to declare tool support, add argument validation with Zod, and handle errors gracefully.
Then progress to exposing structured data through resources and creating reusable prompts that enable AI assistants to interact with your applications intelligently. By the end, you'll have built a complete journaling app that's fully accessible via MCP with discoverable resources and intelligent prompt-based workflows.
Workshop 2: Advanced MCP Features
Days 3-4: Advanced Tools, Elicitation & Long-Running Tasks
Elevate your MCP server with production-grade features. Learn to add sophisticated tool metadata (readOnly, destructive, idempotent annotations) and implement structured output schemas for machine-validated results. Master elicitation techniques that allow your tools to pause workflows and request structured input from users mid-process.
Discover how to use MCP sampling to "borrow" the user's LLM for dynamic content generation. Finally, tackle real-world asynchronous operations with progress reporting, cancellation flows, and real-time notifications that keep clients synchronized with your server's state changes.
Workshop 3: MCP UI
Days 5-6: Simple UI & Advanced UI Features
Transform your MCP server responses from plain text into rich, interactive user interfaces. Start with basic HTML responses using the @mcp-ui/server package and learn how remote-dom ensures consistency with host applications.
Progress to building responsive, interactive UIs with modern frameworks that can display lists, trigger actions, and communicate seamlessly with the host application through post messages. You'll understand the boundary between embedded websites and individual views that enhance rather than replace the conversation experience.
Workshop 4: MCP Auth
Days 7-8: OAuth Setup & Token Resolution
Secure your MCP server with enterprise-grade authentication using the official MCP authorization specification. Learn to implement OAuth 2.1 flows that enable user-specific interactions with your services. Master the process of exposing OAuth metadata, triggering authorization flows from host applications, and resolving client tokens into MCP Auth Info for secure API calls.
Understand how to implement scope-based security to protect your server against unauthorized requests while providing seamless user experiences across different AI platforms and tools.
What's included:
Four complete MCP workshops with tightly edited videos and carefully crafted exercises
Two weeks of live cohort experience working through new material (almost) every day
Lifetime access to all videos and materials in your cohort
An active learning community on Discord
Regular live Q&A sessions with Kent (2x days per week, with accommodations for different timezones)
Completion certificates
30-day money-back guarantee