Screenshot

🚀 Screenshot Server (File Path Focus)

This project offers an MCP (Modular Communication Protocol) server designed to capture screenshots and facilitate their use in other processes or AI assistants, primarily by saving screenshots to the file path specified by the client (host).

🚀 Quick Start

Core Issues and Solutions

Directly interpreting screenshot image data sent by an AI assistant via MCP has proven unreliable in testing. This server adopts a more reliable file path-based workflow:

Recommended Workflow (WSL Host -> Windows Server):

1. An MCP host (such as an AI assistant running in WSL) calls the save_screenshot_to_host_workspace tool and provides its WSL workspace path as a parameter.
2. The server running on Windows captures the screen.
3. The server converts the received WSL path to a Windows-accessible UNC path (e.g., \\wsl$\Distro\path).
4. The server saves the screenshot to the specified location in the host's WSL file system, accessible via the UNC path.
5. The server returns "success" or "failed:... ".
6. The MCP host knows that the screenshot has been saved in its workspace (or in the specified subdirectory if provided in the parameters).
7. The MCP host can pass the WSL path to another dedicated MCP server (running in WSL) for image analysis.

General Alternative Workflow:

1. The MCP host calls take_screenshot_and_return_path, optionally specifying a file name.
2. The server saves the screenshot to its local images/ directory.
3. The server returns the file path.
4. The MCP host can process the screenshot based on the returned path.

✨ Features

Tools

• save_screenshot_to_host_workspace: Saves the screenshot to the host's WSL workspace path.
• take_screenshot_and_return_path: Returns the file path of the screenshot.
• capture_screen: Captures the current screen and returns Base64-encoded image data.

📦 Installation

Setup Instructions

• Install Python on Windows and install project dependencies via uv: uv sync .
• Update the MCP host configuration to start this server:

{
    "mcpServers": {
        "Screenshot-server": {
            "command": "powershell.exe",
            "args": [
                "-Command",
                "cd '<YOUR_WINDOWS_PROJECT_PATH>'; & '<YOUR_WINDOWS_UV_PATH>' run screenshot.py"
            ]
        }
        // ... other servers ...
    }
}

📚 Documentation

File Structure

• screenshot.py: The core MCP server script.
• README.md: This documentation file.
• pyproject.toml: Project definition and dependencies for uv.
• uv.lock: Dependency lock file.
• .gitignore: Git ignore configuration.
• .python-version: (Optional) Python version specifier.
• server.log: Server-generated log file.
• images/: The default directory for take_screenshot_and_return_path.

💻 Usage Examples

Basic Usage

1. In WSL, the AI assistant obtains its current workspace path (e.g., /home/user/current_project).
2. The AI assistant uses use_mcp_tool to call save_screenshot_to_host_workspace on the Screenshot-server, passing host_workspace_path="/home/user/current_project" and an optional name parameter.
3. It receives a "success" response.
4. The AI assistant knows that the screenshot is now located at /home/user/current_project/workspace_screenshot.jpg (or the specified file name).
5. The AI assistant uses use_mcp_tool to call the image analysis server/tool running in WSL, passing the WSL path /home/user/current_project/workspace_screenshot.jpg.
6. The image analysis server reads the file and performs its task.