MCP Screenshot Server

🚀 MCP Screenshot Server

A lightweight, MCP-compatible screenshot microservice built with FastAPI for Windows. It enables AI agents and automation tools to capture full-screen, region-specific, or window-specific screenshots via simple HTTP calls.

🚀 Quick Start

This MCP Screenshot Server is a practical tool that allows you to capture screenshots through straightforward HTTP calls. You can quickly start using it by following the installation steps below.

✨ Features

• 📸 Capture full-screen screenshots.
• 🪟 Capture specific windows by title.
• 🔲 Capture screenshots of a specified region [x, y, width, height].
• 🧠 MCP-compatible REST API.
• 🖼️ Return images in PNG or base64 format.
• 🚀 Built with FastAPI, suitable for production or LLM usage.

📦 Installation

git clone https://github.com/yourusername/mcp-screenshot-server.git
cd mcp-screenshot-server
python -m venv .venv
source .venv/bin/activate   # or .venv\Scripts\activate on Windows
pip install -r requirements.txt
uvicorn main:app --reload

💻 Usage Examples

Basic Usage

POST /screenshot

Request JSON Body:

{
  "region": [0, 0, 1280, 720],        // Optional parameter
  "window_title": "Untitled - Notepad", // Optional parameter
  "base64": true                      // Optional parameter (Default: false)
}

Response (base64 mode):

{
  "status": "ok",
  "mode": "region",
  "image_format": "base64",
  "image": "<base64-encoded-image>"
}

📚 Documentation

Use Cases

• Integration with LLMs via the Model Context Protocol (MCP).
• QA test automation pipelines.
• Monitoring and remote screenshot tools.
• Visual logging/debugging tools for agents.

🔧 Technical Details

Technology Stack

• Python 3.11+
• FastAPI
• mss or pyautogui screenshot tools
• pillow image processing library
• pygetwindow window matching (optional)

📄 License

This project is under the MIT License. You are free to use, fork, and integrate it. For more details, please refer to LICENSE.

📬 Contribution Guidelines

Pull requests and issues are welcome! You can add features or improve cross-platform compatibility (e.g., Mac/Linux support) through PRs.

🙋 Frequently Asked Questions

• Can it run on Linux/macOS? The current version focuses on Windows, but you're welcome to extend it to support other platforms.

• Is it MCP-certified? This project aims to follow the MCP specification as closely as possible to achieve maximum compatibility with LLM agents.

🧠 Inspiration

• Anthropic's Model Context Protocol
• Real-world LLM-driven automation scenarios