Skip to main content

Tool calling using MCP with Semantic Kernel

Semantic Kernel has the concept of a plugin as a way to give your AI superpowers and allow it to perform actions that it wouldn’t be able to do otherwise. A plugin consists of one or more functions that are available to your AI to be invoked.

In Semantic Kernel you can create native plugins that are written in C#, Java or Python or create a plugin based on an OpenAPI specification. Behind the scenes, Semantic Kernel leverages function calling (also called tool calling). With function calling, LLMs can request a particular function, invoke it and capture the results.

The way that these tools are created and called can be different from AI to AI and platform to platform. To tackle this problem Anthropic created the Model Context Protocol as a way to standardize all these integrations.

Think of MCP like a USB-C port for AI applications. Just as USB-C provides a standardized way to connect your devices to various peripherals and accessories, MCP provides a standardized way to connect AI models to different data sources and tools.

The team behind Semantic Kernel didn’t hesitate and added support for MCP based plugins as well (currently in preview). Let me show you how to use it…

Using an MCP server as a plugin in Semantic Kernel

The first thing you need is an MCP server. Today there are typically 2 ways that an MCP server can exchange information:

  • Standard Input Output (STDIO)
  • Server Send Events(SSE)

I will be using SSE in this example with the Playwright MCP server.

  • Start the server by executing the following command:

npx @playwright/mcp@latest --port 8931

  • Now open your project(I assume you already have the basic Semantic Kernel setup done) and add the following NuGet package:

dotnet add ModelContextProtocol-SemanticKernel

  • Next step is to load the tools provided by the MCP server as a plugin:
  • Last important step is that we enable automatic function invocation on the AI model so that it can invoke the available tools.
  • Now we can interact with our AI model as before:

That’s all!

For an end-to-end demo, check out my Semantic Kernel demo application: wullemsb/SemanticKernel at MCP

More information

Introduction - Model Context Protocol

Semantic Kernel – Auto function calling

wullemsb/SemanticKernel at MCP

microsoft/playwright-mcp: Playwright MCP server

Plugins in Semantic Kernel | Microsoft Learn

Let GitHub Copilot interact with your local PowerPoint and Word documents

Labels:

Popular posts from this blog

Podman– Command execution failed with exit code 125

After updating WSL on one of the developer machines, Podman failed to work. When we took a look through Podman Desktop, we noticed that Podman had stopped running and returned the following error message: Error: Command execution failed with exit code 125 Here are the steps we tried to fix the issue: We started by running podman info to get some extra details on what could be wrong: >podman info OS: windows/amd64 provider: wsl version: 5.3.1 Cannot connect to Podman. Please verify your connection to the Linux system using `podman system connection list`, or try `podman machine init` and `podman machine start` to manage a new Linux VM Error: unable to connect to Podman socket: failed to connect: dial tcp 127.0.0.1:2655: connectex: No connection could be made because the target machine actively refused it. That makes sense as the podman VM was not running. Let’s check the VM: >podman machine list NAME         ...
Read more

Cache stampede: when our cache turned against us

While investigating some performance issues, we ran into an ASP.NET Core API that cached a fairly expensive aggregation query for 60 seconds. Under normal load, that was fine: one request rebuilds the cache, everyone else reads from it. Under peak load, dozens of requests would arrive in that same expiry window, all see a cache miss, and all fire the same expensive query in parallel. The database didn't like that. That was the moment when our caching layer stopped helping and started hurting. A burst of requests comes in at the same time, all miss the cache, and all go hammer the database or the downstream API at once. That's a cache stampede . The cache was supposed to protect our backend, and for a few hundred milliseconds it did the opposite. Why this happens IMemoryCache.GetOrCreate (and its async sibling) looks like it protects you, but it doesn't add any locking on its own. Look at the naive version: public async Task<Report> GetReportAsync(string key) ...
Read more

Azure DevOps/ GitHub emoji

I’m really bad at remembering emoji’s. So here is cheat sheet with all emoji’s that can be used in tools that support the github emoji markdown markup: All credits go to rcaviers who created this list.
Read more