Some experiments around AI to learn.
- MCPSSEServer Simple .NET C# MCP Server using SSE transport using modelcontextprotocol / csharp-sdk published using AOT in a single Windows exe of 15.7MB!
- MCPServerDocker Two simple console applications using
modelcontextprotocol / csharp-sdk, one .NET C# MCP Server and its client, and a way to publish the server as a Docker image. - MCPServer Two simple console applications using modelcontextprotocol / csharp-sdk, Microsoft.Extensions.AI with Ollama & tools to build a Model Context Protocol (MCP) server and client!
- OllamaMCPServerMicrosoftExtensions simple console application using Microsoft.Extensions.AI with Ollama & tools from a Model Context Protocol (MCP) server!
- SKOllamaAgentWithFunction demonstrates the way to create a Semantic Kernel Agent with function in C# using Ollama to run local SLM model like llama3.2.
- SKOllamaAgent demonstrates the way to create a Semantic Kernel Agent in C# using Ollama to run local SLM model like phi4.
- DeepSeekOllamaAspire explore how to run DeepSeek-R1 distilled model by harnessing the capabilities of .NET Aspire alongside Ollama on your local environment.
- OllamaToolCallingMicrosoftExtensions a discovery project to learn how AI tool/function calling are working using Microsoft.Extensions.AI. Follows my own implementation OllamaToolConsoleApp.
- OllamaToolCallingOpenAPI a console application using Semantic Kernel and its OpenAPI plugin to call an ASP.NET Core minimal API, using the new .NET 9 OpenAPI support, all on your local machine with Ollama.
- OllamaToolConsoleApp a discovery project to learn how AI tool/function calling are working. A C# source generator is used to annotate a C# method which can then be called by Ollama using its tool support.
- Phi3SKConsoleApp a console application that interacts with Phi-3 on your machine using Ollama, C# and Semantic Kernel.
- Phi3SKWebApp A blazor web application leveraging Fluent UI Blazor components library that interacts with Phi-3 on your machine using Ollama, C# and Semantic Kernel. It demonstrates the streaming of chat completion from Phi-3 to the web application in real-time.
From my .NET Aspire playground
- LocalLLMSummarize a simple web application which get a saved article from omnivore.app and summarize the articles using a local LLM and display the summary to the user.
- OpenAI a simple web application which get a saved article from omnivore.app and summarize the articles using Azure OpenAI and display the summary to the user.