As the list of available tools keeps growing, sooner or later something will not work and some debugging becomes necessary. In this post I look into ways to troubleshoot your MCP integration. Let's dive in! To understand what we need to debug you need to be aware of the architecture of an MCP integration. It follows a client-server architecture , where AI models can request data through a host with an MCP client (e.g. GitHub Copilot, Claude Desktop, …) from MCP servers, which then retrieve relevant information from local or remote sources. This means that when a problem occurs that there (at least) 2 places to look at. Debugging the MCP client Let’s start by looking into the client. The way you need to debug the client is completely dependent on the host. I’ll focus on GitHub Copilot and Claude Desktop. GitHub Copilot When VSCode encounters an error while trying to interact with an MCP server, you get a red error indicator in the Chat window: Click on the icon and...
On Monday I talked about interacting with Word and Powerpoint documents directly through GitHub Copilot. After writing that post , a colleague reached out to me asking if the same is possible through Claude Desktop . As the original MCP protocol was created by Anthropic(the company that created Claude Desktop), the answer is 'of course'. Remark: If you want a general introduction about MCP servers, check out my previous post . The use case Part of my job is working on presales activities. Typically, this involves answering RFI's and RFP's by creating solution designs, project plans and describe the supporting processes. The end result is typically a Word document containing all the details about our offer in combination with a PowerPoint to present our offer to the customer. I was wondering if I could let an AI agent interact directly with these documents to help me create and finetune our proposals. Use an MCP server to talk to Microsoft Office To realize th...