ScreenMonitorMCP

ScreenMonitorMCP v2 is an MCP server that enables an AI assistant to interact with and understand your screen content in real time. It provides tools to capture screenshots, analyze screen data with AI models, analyze individual images, and stream live screen content, all while reporting system performance metrics. With these capabilities, your AI can help with UI/UX analysis, debugging assistance, content documentation, accessibility checks, and overall visual system monitoring. To use it, configure the MCP client (for example Claude Desktop) to point at the ScreenMonitorMCP v2 entry point using the provided Python module invocation. The server exposes tools such as capture_screen, analyze_screen, analyze_image, create_stream, and get_performance_metrics that you can invoke through the MCP client to perform tasks like grabbing a screenshot, performing image-based analysis, or starting a live video stream of your display.

Prerequisites:

• Python 3.8 or newer
• pip (comes with Python)
• Access to install Python packages from PyPI

Installation steps:

1. Install from PyPI:

pip install screenmonitormcp

2. Alternatively, install from source:

git clone https://github.com/inkbytefo/screenmonitormcp.git
cd screenmonitormcp
pip install -e .

3. Run the MCP server module directly (example):

python -m screenmonitormcp_v2.mcp_main

4. Configure your MCP client (e.g., Claude Desktop) to use the Python module entry point as shown in the README configuration snippet. Ensure you provide your OpenAI API credentials and model settings via environment variables as needed.

Environment and configuration tips:

• Set OPENAI_API_KEY to a valid API key for your OpenAI-compatible service.
• OPENAI_BASE_URL can be used to point to alternative endpoints if you’re not using the default OpenAI service.
• OPENAI_MODEL should reference the desired model/version compatible with your setup.
• Ensure your MCP client supports Python-based MCP servers and can mount environment variables for the server process.
• If you’re debugging, start with capture_screen to verify screenshot capture works, then move to analyze_screen for AI-assisted analysis.
• For multi-monitor setups, confirm the MCP client and OS permissions allow screen capture across all displays.
• Keep dependencies up-to-date to avoid compatibility issues with newer vision models or OpenAI API changes.