
Discover how the Model Context Protocol (MCP) standardizes ai-model interactions with external tools via JSON-RPC over websockets, enabling secure, real-time access to data sources.
Discover how to build a server sent events system in .NET, push real-time updates over a single persistent http connection, and implement a unidirectional server and client with content-type text/event-stream.
Create a dotnet server with http listener on port 8888 that streams server-sent events to the stream path, with logging. Use text event stream and no-cache; send every two seconds.
Create a long-lived http client to receive server-sent events, disable buffering, and stream data line by line, reading with a reader and printing each line to the console.
Test the SSH server and client workflow by running Linqpad server on port 8888 and connecting a client to the stream endpoint to read push messages and handle client streams.
Explore how MCP uses SEC in a smart AI app, orchestrating a music service via prompts; the LLM selects a tool, then MCP client–server handles the call and results.
Build a minimal MCP server from scratch that executes shell commands via JSON RPC, supports a single client, and streams real-time responses using server-sent events.
Build an MCP client from scratch, set up an HTTP client, handshake via server-sent events, and use task completion sources with cancellation tokens to send JSON RPC shell commands.
Test the MCP server and client interaction by starting the server, initializing the connection, and sending shell requests. Receive JSON RPC over HTTP and streamed responses via server sent events.
Add AI reasoning to MCP client by integrating LLMs and semantic kernel to decide which tool to invoke, manage chat history, and test shell command outputs for production readiness.
Explore how the model context protocol (MCP) in C# enables LLMs to adapt and explore server resources in real time, guiding application workflows and LLM orchestration.
Discover building a basic MCP server in a console app using NuGet and dependency injection, exposing name and version, with echo and sum tools.
Create a basic MCP client with the MCP SDK, connect to a manually started server, and explore available tools like echo and sum, using dotnet run for the C# server.
Explore integrating MCP Tools as semantic kernel functions in C#, converting MCP server tools into semantic kernel calls, with logging, OpenAI prompts, and testing in Visual Studio.
Orchestrate Semantic Kernel agents in a sequential workflow to read text from a file, translate it, and log the process using kernels, execution settings, and a chat history.
Explore concurrent orchestration of technologist and economist agents using an OpenAI kernel, with logging, a history object, and a collector to derive top ideas for the maritime industry.
Orchestrate a group chat using semantic kernel tools by creating code generator and runner agents, maintaining a conversation history, and producing a hello world in C#.
Learn to connect a dotnet MSSQL MCP server to a SQL Server database using an environment variable for the connection string, enabling the coding agent to manage tables.
Learn how to build a minimal Model Context Protocol (MCP) server and client using C#. This hands-on course is perfect for developers who want to understand how LLMs like Claude interact with tools and services using the MCP protocol.
We’ll start from scratch and walk through creating a simple SSE(Server Sent Events) MCP server.
Then, we’ll build a lightweight client that connects and talks to this server.
Next, we will upgrade our client and server to conform to MCP standard using JSON-RPC and SSE communication.
Finally we will add AI part to our application using Semantic Kernel to fully understand why MCP was created and how it works.
You'll understand the full request-response flow, the structure of MCP messages, and how client interacts with the server.
No advanced setup required — you can code along each lesson using the links provided, or using your preferred IDE.
By the end of the course, you'll have a working MCP example and the knowledge to build more advanced integrations for your own tools, bots, or AI services.
What You’ll Learn:
How MCP works and why it matters
Building an MCP-compliant server using SSE in C# from scratch
Sending JSON-RPC messages from a minimal client
Handling requests like initialize, resources, and tools
Best practices and next steps toward full compliance
Using MCP with Semantic Kernel
Using latest MCP Servers in your coding workflow
Who This Course Is For:
C# developers curious about LLM integrations
Beginners learning protocol design
Anyone building tools or services that connect to AI via MCP