Skip to main content

Manual

MCP server overview​

With the Multiplayer MCP (Model Context Protocol), you can feed your coding agents session-based, unsampled runtime data, pre-correlated across your entire stack: from frontend screens and user actions, to backend traces, metrics, logs, and full request/response content and headers for every component in your system.

Your AI tools can generate accurate fixes, tests, and features using pre-correlated and accurately scoped data.

Prerequisites​

Follow the Multiplayer quickstart.

Manual setup overview​

Choose the configuration method that works best for your editor.

For clients that support OAuth, use this streamlined configuration:

{
  "mcpServers": {
    "Multiplayer": {
      "url": "https://api.multiplayer.app/v0/api/public/mcp"
    }
  }
}

Remote MCP configuration​

For clients that don't support OAuth, use this configuration:

{
  "mcpServers": {
    "Multiplayer": {
      "command": "npx",
      "args": ["-y", "mcp-remote@latest", "https://api.multiplayer.app/v0/api/public/mcp"]
    }
  }
}

Next steps​

🚀 If you’re ready to trial Multiplayer with your own app, you can follow the steps in the quickstart.

📌 If you have any questions or want to report bugs please open a GitHub issue! 💜