Peggle AI

Provides an MCP endpoint that captures Peggle screen data and issues mouse clicks to play the game automatically.

javascript

0

GitHub Stars

javascript

Language

6 months ago

First Indexed

3 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "collabvmgamez-peggle-ai-mcp": {
      "command": "node",
      "args": [
        "C:\\Users\\{youruser}\\peggle-ai-mcp\\dist\\index.js"
      ]
    }
  }
}

This MCP server lets an AI play Peggle by capturing the screen and issuing mouse clicks, enabling automated play through MCP clients. It uses vision to analyze the game screen and then sends precise mouse actions to select shots.

How to use

You integrate this server with an MCP client to control Peggle automatically. Start the server from your local environment, then connect through an MCP client tool (such as LM Studio) to issue tasks like taking a screenshot, analyzing the board, and clicking the optimal shot.

Typical usage flow you perform in your MCP client:\n- Start Peggle on your primary monitor.\n- Launch the MCP client and add this server as an available endpoint.\n- Choose a model that supports vision and tools.\n- Instruct the AI to take a screenshot, analyze the board, and click the best shot. The server will respond by capturing the screen, processing the image, and performing the click at the chosen coordinates.

How to install

Prerequisites: ensure you have Node.js and npm installed on your machine.

npm install
npm run build

Configure LM Studio to run this MCP server. In the MCP tab, add a new server and choose the stdio configuration. Provide the following runtime details exactly as shown.

Command: node
Arguments: C:\Users\{youruser}\peggle-ai-mcp\dist\index.js

If you prefer using npx, you can run the server with npx by pointing LM Studio to the built package location.

Command: npx
Arguments: -y C:\Users\maxwe\Desktop\peggle-ai-mcp

StartPeggle workflow in LM Studio by opening Peggle on your primary monitor, selecting a vision-enabled model, and instructing the AI to take a screenshot, analyze the best shot, and click there.

Implementation notes and troubleshooting

Screen capture is implemented with a cross-platform screen capture module, and mouse control is performed through PowerShell commands invoked from the server for compatibility on Windows without native build tools.

If you run into Windows 11 PowerShell execution policy issues, adjust the policy to allow script execution. The server uses standard PowerShell calls that work under typical default configurations.

Notes with Windows and environment considerations

The server is configured to operate with a single primary path for the built index.js when started via Node. Ensure Peggle is visible on your primary monitor for accurate screen capture and click targeting.

Available tools

capture_screen

Captures the primary monitor and returns the image as base64 for analysis by the model.

click_at

Moves the mouse to (x, y) and performs a left click to execute a shot in Peggle.