OmniParser AutoGUI MCP server

This MCP server analyzes your screen using OmniParser and automates GUI interactions based on those insights. It acts as a bridge between AI assistants and your computer's interface, allowing for automated actions like clicking buttons, entering text, or navigating applications.

Installation

Prerequisites

You'll need Git and Python with uv installed on your system. This server has been confirmed to work on Windows.

Setup Steps

1. Clone the repository with submodules:

   git clone --recursive https://github.com/NON906/omniparser-autogui-mcp.git
   cd omniparser-autogui-mcp
   

2. Install dependencies and download required models:

   uv sync
   

   For Windows:

   set OCR_LANG=en
   uv run download_models.py
   

   For Linux/macOS:

   export OCR_LANG=en
   uv run download_models.py
   

   If you plan to use the LangChain example:

   uv sync --extra langchain
   

Configuration

Add the MCP server to your Claude Desktop configuration by editing the claude_desktop_config.json file:

{
  "mcpServers": {
    "omniparser_autogui_mcp": {
      "command": "uv",
      "args": [
        "--directory",
        "D:\\YOUR_PATH\\omniparser-autogui-mcp",
        "run",
        "omniparser-autogui-mcp"
      ],
      "env": {
        "PYTHONIOENCODING": "utf-8",
        "OCR_LANG": "en"
      }
    }
  }
}

Be sure to replace D:\\YOUR_PATH\\omniparser-autogui-mcp with the actual path where you cloned the repository.

Environment Configuration

The server supports several environment variables to customize its behavior:

Core Settings

• OCR_LANG: Language for OCR (default: "en")
• TARGET_WINDOW_NAME: Specifies a particular window to operate on (if not set, operates on the entire screen)
• OMNI_PARSER_BACKEND_LOAD: Set to 1 if using with clients other than Claude Desktop (like LibreChat)

Advanced Configuration

• OMNI_PARSER_SERVER: Address of a separate OmniParser server (e.g., "127.0.0.1:8000")
• SSE_HOST and SSE_PORT: Enable Server-Sent Events (SSE) communication instead of stdio
• SOM_MODEL_PATH: Custom path for the SOM model
• CAPTION_MODEL_NAME: Custom caption model name
• CAPTION_MODEL_PATH: Custom path for the caption model
• OMNI_PARSER_DEVICE: Device to run OmniParser on
• BOX_TRESHOLD: Detection threshold

Usage

Once installed and configured, the MCP server can perform various GUI operations based on visual analysis:

• Identify and interact with UI elements (buttons, text fields, checkboxes)
• Read text from the screen
• Enter text into forms
• Navigate through applications
• Perform clicks and other mouse actions

Example Tasks

• Search for specific text in a browser
• Fill out forms based on instructions
• Find and click buttons based on their appearance or surrounding context
• Extract information from visual elements on screen

The server works in conjunction with Claude or other AI assistants that support the Model Context Protocol.

How to add this MCP server to Cursor

There are two ways to add an MCP server to Cursor. The most common way is to add the server globally in the ~/.cursor/mcp.json file so that it is available in all of your projects.

If you only need the server in a single project, you can add it to the project instead by creating or adding it to the .cursor/mcp.json file.

Adding an MCP server to Cursor globally

To add a global MCP server go to Cursor Settings > MCP and click "Add new global MCP server".

When you click that button the ~/.cursor/mcp.json file will be opened and you can add your server like this:

{
    "mcpServers": {
        "cursor-rules-mcp": {
            "command": "npx",
            "args": [
                "-y",
                "cursor-rules-mcp"
            ]
        }
    }
}

Adding an MCP server to a project

To add an MCP server to a project you can create a new .cursor/mcp.json file or add it to the existing one. This will look exactly the same as the global MCP server example above.

How to use the MCP server

Once the server is installed, you might need to head back to Settings > MCP and click the refresh button.

The Cursor agent will then be able to see the available tools the added MCP server has available and will call them when it needs to.

You can also explictly ask the agent to use the tool by mentioning the tool name and describing what the function does.