The official C# SDK for the Model Context Protocol, enabling .NET applications, services, and libraries to implement and interact with MCP clients and servers. Please visit our API documentation for more details on available functionality.
This SDK consists of three main packages:
ModelContextProtocol
- The main package with hosting and dependency injection extensions. This is the right fit for most projects that don't need HTTP server capabilities. This README serves as documentation for this package.
ModelContextProtocol.AspNetCore
- The library for HTTP-based MCP servers. Documentation
ModelContextProtocol.Core
- For people who only need to use the client or low-level server APIs and want the minimum number of dependencies. Documentation
Note
This project is in preview; breaking changes can be introduced without prior notice.
The Model Context Protocol (MCP) is an open protocol that standardizes how applications provide context to Large Language Models (LLMs). It enables secure integration between LLMs and various data sources and tools.
For more information about MCP:
To get started, install the package from NuGet
dotnet add package ModelContextProtocol --prerelease To get started writing a client, the McpClient.CreateAsync method is used to instantiate and connect an McpClient to a server. Once you have an McpClient, you can interact with it, such as to enumerate all available tools and invoke tools.
usingModelContextProtocol.Client;usingModelContextProtocol.Protocol;varclientTransport=newStdioClientTransport(newStdioClientTransportOptions{Name="Everything",Command="npx",Arguments=["-y","@modelcontextprotocol/server-everything"],});varclient=awaitMcpClient.CreateAsync(clientTransport);// Print the list of tools available from the server.foreach(vartoolinawaitclient.ListToolsAsync()){Console.WriteLine($"{tool.Name} ({tool.Description})");}// Execute a tool (this would normally be driven by LLM tool invocations).varresult=awaitclient.CallToolAsync("echo",newDictionary<string,object?>(){["message"]="Hello MCP!"},cancellationToken:CancellationToken.None);// echo always returns one and only one text content objectConsole.WriteLine(result.Content.OfType<TextContentBlock>().First().Text);You can find samples demonstrating how to use ModelContextProtocol with an LLM SDK in the samples directory, and also refer to the tests project for more examples. Additional examples and documentation will be added as in the near future.
Clients can connect to any MCP server, not just ones created using this library. The protocol is designed to be server-agnostic, so you can use this library to connect to any compliant server.
Tools can be easily exposed for immediate use by IChatClients, because McpClientTool inherits from AIFunction.
// Get available functions.IList<McpClientTool>tools=awaitclient.ListToolsAsync();// Call the chat client using the tools.IChatClientchatClient= ...;varresponse=awaitchatClient.GetResponseAsync("your prompt here",new(){Tools=[.. tools]},Here is an example of how to create an MCP server and register all tools from the current application. It includes a simple echo tool as an example (this is included in the same file here for easy of copy and paste, but it needn't be in the same file... the employed overload of WithTools examines the current assembly for classes with the McpServerToolType attribute, and registers all methods with the McpServerTool attribute as tools.)
dotnet add package ModelContextProtocol --prerelease dotnet add package Microsoft.Extensions.Hosting usingMicrosoft.Extensions.DependencyInjection;usingMicrosoft.Extensions.Hosting;usingMicrosoft.Extensions.Logging;usingModelContextProtocol.Server;usingSystem.ComponentModel;varbuilder=Host.CreateApplicationBuilder(args);builder.Logging.AddConsole(consoleLogOptions =>{// Configure all logs to go to stderrconsoleLogOptions.LogToStandardErrorThreshold=LogLevel.Trace;});builder.Services.AddMcpServer().WithStdioServerTransport().WithToolsFromAssembly();awaitbuilder.Build().RunAsync();[McpServerToolType]publicstaticclassEchoTool{[McpServerTool,Description("Echoes the message back to the client.")]publicstaticstringEcho(stringmessage)=>$"hello {message}";}Tools can have the McpServer representing the server injected via a parameter to the method, and can use that for interaction with the connected client. Similarly, arguments may be injected via dependency injection. For example, this tool will use the supplied McpServer to make sampling requests back to the client in order to summarize content it downloads from the specified url via an HttpClient injected via dependency injection.
[McpServerTool(Name="SummarizeContentFromUrl"),Description("Summarizes content downloaded from a specific URI")]publicstaticasyncTask<string>SummarizeDownloadedContent(McpServerthisServer,HttpClienthttpClient,[Description("The url from which to download the content to summarize")]stringurl,CancellationTokencancellationToken){stringcontent=awaithttpClient.GetStringAsync(url);ChatMessage[]messages=[new(ChatRole.User,"Briefly summarize the following downloaded content:"),new(ChatRole.User,content),];ChatOptionsoptions=new(){MaxOutputTokens=256,Temperature=0.3f,};return$"Summary: {awaitthisServer.AsSamplingChatClient().GetResponseAsync(messages,options,cancellationToken)}";}Prompts can be exposed in a similar manner, using [McpServerPrompt], e.g.
[McpServerPromptType]publicstaticclassMyPrompts{[McpServerPrompt,Description("Creates a prompt to summarize the provided message.")]publicstaticChatMessageSummarize([Description("The content to summarize")]stringcontent)=>new(ChatRole.User,$"Please summarize this content into a single sentence: {content}");}More control is also available, with fine-grained control over configuring the server and how it should handle client requests. For example:
usingModelContextProtocol;usingModelContextProtocol.Protocol;usingModelContextProtocol.Server;usingSystem.Text.Json;McpServerOptionsoptions=new(){ServerInfo=newImplementation{Name="MyServer",Version="1.0.0"},Handlers=newMcpServerHandlers(){ListToolsHandler=(request,cancellationToken)=>ValueTask.FromResult(newListToolsResult{Tools=[newTool{Name="echo",Description="Echoes the input back to the client.",InputSchema=JsonSerializer.Deserialize<JsonElement>("""{ "type": "object", "properties":{ "message":{ "type": "string", "description": "The input to echo back" } }, "required": ["message"] } """),}]}),CallToolHandler=(request,cancellationToken)=>{if(request.Params?.Name=="echo"){if(request.Params.Arguments?.TryGetValue("message",outvarmessage)is not true){thrownewMcpProtocolException("Missing required argument 'message'",McpErrorCode.InvalidParams);}returnValueTask.FromResult(newCallToolResult{Content=[newTextContentBlock{Text=$"Echo: {message}",Type="text"}]});}thrownewMcpProtocolException($"Unknown tool: '{request.Params?.Name}'",McpErrorCode.InvalidRequest);}}};awaitusingMcpServerserver=McpServer.Create(newStdioServerTransport("MyServer"),options);awaitserver.RunAsync();Descriptions can be added to tools, prompts, and resources in a variety of ways, including via the [Description] attribute from System.ComponentModel. This attribute may be placed on a method to provide for the tool, prompt, or resource, or on individual parameters to describe each's purpose. XML comments may also be used; if an [McpServerTool], [McpServerPrompt], or [McpServerResource]-attributed method is marked as partial, XML comments placed on the method will be used automatically to generate [Description] attributes for the method and its parameters.
The starting point for this library was a project called mcpdotnet, initiated by Peder Holdgaard Pedersen. We are grateful for the work done by Peder and other contributors to that repository, which created a solid foundation for this library.
This project is licensed under the MIT License.
