Skip to main content

How connecting works

Every MCPCore server has a permanent subdomain endpoint:
https://your-subdomain.mcpcore.io/mcp
You copy the ready-to-use config snippet from the server’s Integration tab and paste it into your AI client’s config file. The client connects, discovers your tools, and is ready to use them.
Integration tab showing client selector and config snippet

Supported clients


Config formats at a glance

Most clients use a command-based config that runs mcp-remote to bridge HTTP/SSE:
{
  "mcpServers": {
    "your-server-code": {
      "command": "npx",
      "args": ["-y", "mcp-remote@latest", "https://your-subdomain.mcpcore.io/mcp"]
    }
  }
}
Cursor and VS Code support native HTTP transport and don’t need npx:
{
  "mcpServers": {
    "your-server-code": {
      "url": "https://your-subdomain.mcpcore.io/mcp"
    }
  }
}

Auth headers by security mode

Security modeWhat to add to the config
PublicNothing extra
UnlistedUse the full URL with the embedded key
API KeyAdd --header "Authorization: Bearer <key>" (command) or headers.Authorization (URL-based)
OAuthNot yet supported in all clients — check client-specific docs

The Integration tab

The easiest way to get the right config is to use the Integration tab on your server’s detail page. Select your client, and the tab generates the correct snippet with your subdomain and API key pre-filled — just copy and paste.
Integration tab with client selector and copy button