Hey everyone! I'm exploring the use of Claude Code as an MCP server, but I'm running into some confusion. The documentation doesn't provide enough detail. I want to set up two Claude MCP instances—one for my backend and another for my frontend. My goal is to connect them to my Claude desktop app for management. Has anyone successfully pulled this off? Any tips would be greatly appreciated!
1 Answer
Just a heads up, Claude Code acts more like an MCP client rather than a server. There’s a significant difference here. The MCP server is the one that holds the tools and processes requests, while the client—like Claude Code or the desktop app—interacts with the server. I think there might be some confusion regarding these roles!
Thanks for clarifying that! But I noticed that you can connect Claude Code to another MCP server using settings from the docs. Have you checked that out?