request a example project

[Salvions]

Pre-submission Checklist

• I have checked that this question would not be more appropriate as an issue in a specific repository
• I have searched existing discussions and documentation for answers

Question Category

• Protocol Specification
• SDK Usage
• Server Implementation
• General Implementation
• Documentation
• Other

Your Question

there is an open-source examploe project who implements a client and a server i two different solution?

[mikekistler]

The samples directory of this repo has several sample clients and servers. Is that what you are looking for?

[StefanBelo]

Did any code or samples actually go through unit testing? The AI agent tools execution isn't working, and I couldn't find how WithSubscribeToResourcesHandler was tested. More importantly, how do I make an AI agent client subscribe to an MCP resource?

[mikekistler]

@StefanBelo Yes we have tested the samples but of course it's quite possible that we missed things. Please open issues for any bugs you find.

[StefanBelo]

I do not have an OpenAI subscription, but I have a DeepSeek one. Since the DeepSeek client is OpenAI-compatible, I added some code to recreate your sample ChatWithTools.

I think in your sample code, there is no need to use a sampling client just to initiate a chat with tools—correct me if I am wrong.

My MCP Server uses the SSE protocol. In my screenshot, you can see that the AI client with DeepSeek works—when prompted "Who are you?", it replies. You can also see that the MCP client works; I was able to list my MCP server tools.

However, when prompting a question about where the AI agent should reply with tool usage and report what data it gets from the MCP tool, there is no reply. I can see that five calls to the MCP server are made, but no usable reply is returned. It does not matter if sampling is added or not the same reply, with or without sampling, calls to mcp tools are made, no reply.

I conducted the same experiment with FastAgent, the Python package. There, everything works fine with DeepSeek as the AI model and my MCP SSE server.

My test code:

https://github.com/StefanBelo/BetfairAiTrading/tree/main/src/AiAgentCSharp

https://github.com/StefanBelo/BetfairAiTrading/tree/main/src/AiAgentPython

[peidaqi]

Here's a minimal client/server implementation using both stdio and http:

https://github.com/peidaqi/mcp-csharp-minimal

Hope this helps.