Model Context Protocol (MCP) – Multi-language

Stop building brittle, one-off integrations. Whether you're connecting a legacy Java database to Copilot ororchestrating complex C# agents, the Model Context Protocol (MCP) is the universal standard. Write yourintegration once, and let every AI assistant use it.

Do you have an existing application, API, or service that you want to make accessible to AI assistants? Add an MCP endpoint and tools like VS Code Copilot, Claude Desktop, or custom AI agents can instantly use your capabilities – without building custom integrations for each one. Are you orchestrating AI agents that need to discover and invoke tools at runtime? Looking for a standardized way to connect your agents to the world?


Let's standardize with MCP!


Anthropic created the Model Context Protocol (MCP) as an open standard to solve the fragmentation
problem in AI integrations. Since its release in November 2024, MCP has been adopted by Microsoft,
JetBrains, and the broader AI ecosystem. It provides a universal way to connect AI assistants to the tools
and data they need.

This course uses the official MCP SDKs maintained by Anthropic and partners (Microsoft for C#, JetBrains for Kotlin, Spring AI for Java).

What you'll build: In this course, you'll create MCP servers that expose your functionality to AI assistants,
and MCP clients that orchestrate multiple servers. By the end, you'll have a working server connected to VS Code Copilot and understand how to secure it for production use. Students choose their preferred language: TypeScript, C#, Python, or Java/Kotlin. All concepts are demonstrated across languages with complete solutions provided. The course is continuously updated to reflect the latest MCP specification and SDK releases.

Course objectives

  • Expose your applications to AI – Turn any API or service into an MCP server that AI assistants can
    use
  • Build AI applications – Create clients that connect to MCP servers and orchestrate multiple tools
  • Understand the protocol – Learn how MCP works under the hood so you can debug and extend it
  • Secure your servers – Implement authentication and authorization for production deployments
  • Integrate with IDEs – Connect your servers to VS Code Copilot and other AI tools

Prerequisites

Professional developers familiar with at least one supported language (TypeScript, C#, Python, Java, or
Kotlin) and comfortable with JSON, HTTP APIs, and basic asynchronous programming patterns.

Target audience

This course is suitable for:

  • Developers working with AI integrations
  • Software developers building APIs and platform services
  • Architects looking for a standardized way to connect AI to systems
  • Technologists working with agent-based and AI-driven solutions

Rick Beerendonk - instructor of the course

30+ years software development experience, senior consultant at Oblicum, Microsoft MVP. Has trained thousands of developers across Europe in .NET, modern web development, API design, and software architecture.

Introduction to MCP

  • What is MCP and what problems does it solve?
  • The MCP architecture: Hosts, Clients, and Servers
  • Transports: Stdio (local/desktop) vs. SSE (remote/production)
  • Hands-on: Explore MCP messages with MCP Inspector

Building Your First Server

  • Setting up your development environment
  • Creating tools that AI assistants can call
  • Input validation and error handling
  • Hands-on: Build a multi-tool utility server

Exposing Data with Resources

  • Making files, databases, and APIs available to AI
  • Static and dynamic resources
  • Change notifications
  • Hands-on: Build a document server

Creating Reusable Prompts

  • Prompt templates with parameters
  • When to use prompts vs. tools

Sampling: Server-Initiated LLM Requests

  • Enabling "human in the loop" workflows
  • When and how servers can request LLM completions

Building MCP Clients

  • Connecting to MCP servers programmatically
  • Discovering and invoking tools
  • Working with multiple servers
  • Hands-on: Build an interactive MCP client

Logging and Progress

  • Protocol-level logging from your server
  • Progress reporting for long operations

Security Architecture

  • Authentication for remote (SSE) servers
  • Gateway patterns and token validation
  • Input validation and sanitization
  • Hands-on: Secure your remote MCP server

IDE and Agent Integration

  • Connecting servers to VS Code Copilot
  • Testing and debugging workflows
  • Microsoft Semantic Kernel integration (C#/Python)
  • IntelliJ integration (Java/Kotlin)
  • Hands-on: Connect your server to VS Code Copilot

Appendix


Appendix A: Debugging MCP Servers

  • Using MCP Inspector for troubleshooting
  • Common issues and solutions
  • IDE debugging techniques

Appendix B: Testing MCP Servers

  • Unit and integration testing strategies
  • Language-specific testing frameworks
  • Contract testing for JSON-RPC

Practical information

Duration: 3 days
Price: 20.900 kr
Language: English
Format: Can be delivered as an open course or as an in-house course

FAQ

Hvilke programmeringsspråk kan jeg bruke på kurset?
Du kan velge mellom TypeScript, C#, Python eller Java og Kotlin, og alle konsepter demonstreres på tvers av språk.

Er dette et sertifiseringskurs?
Nei, kurset gir ingen formell sertifisering, men gir solid praktisk kompetanse på MCP.

Hvor praktisk er kurset?
Kurset er workshop-basert med hands-on øvelser, eksempler og komplette løsningsforslag.

Må jeg ha erfaring med AI fra før?
Det er en fordel, men ikke et krav. Kurset forklarer hvordan MCP brukes i AI-kontekst fra grunnen av.

Er innholdet oppdatert på nyeste MCP-spesifikasjon?
Ja, kurset oppdateres fortløpende i takt med endringer i MCP-spesifikasjonen og SDK-ene.

Relevant courses

Other subject areas

Andre relevante kurs

AZ-2005: Develop AI agents using Azure OpenAI and the Semantic Kernel SDK
1 dager
Classroom Virtual
AI-102: Develop AI solutions in Azure
20. april
5 dager
Classroom Virtual Startgaranti