1MCP: Run All Your AI Assistants Through One Server 🚀

[xizhibei]

Pre-submission Checklist

• This post follows the MCP community guidelines

What would you like to share?

Tired of running same multiple MCP servers for Claude Desktop, Cursor, and other AI assistants? 1MCP lets you run everything through a single unified server!

⚠️ Currently in Beta - your feedback and bug reports are extremely valuable!

Quick Setup (30 seconds) ⚡️

# Want to use your existing Claude Desktop config? Just add:
npx -y @1mcp/agent --config ~/Library/Application\ Support/Claude/claude_desktop_config.json # For MacOS
npx -y @1mcp/agent --config %APPDATA%\Claude\claude_desktop_config.json # For Windows

Now you can add 1mcp to your AI assistants, such as Cursor (~/.cursor/mcp.json):

{
    "mcpServers": {
        "1mcp": {
            "type": "http",
            "url": "http://localhost:3050/sse"
        }
    }
}

That's it! Now you have:

• 🔄 One server for all AI assistants
• 💪 Better system performance
• 🎯 Smart tool routing between clients
• ⚙️ Simple configuration

Try it out and let me know what you think!

Relevant Links

https://github.com/1mcp-app/agent

[ukpatell]

Just an open question - what do the performance metrics look like? Any testing done for this to handle multiple i.e. thousands / millions connections?

[xizhibei]

This is a great question. At present, the purpose of creating this project is mainly to solve the problem that the same MCP server needs to be configured in multiple places on our local computer, the connections will be limited.

The problem you mentioned seems to be encountered in the cloud, so I did not do this kind of test. However, may I ask you about your specific requirements? Perhaps I can expand the application scenarios of my project.