macOS Screen MCP Server

Example Use Cases

• 1

  Capture a screenshot of a specific macOS application window (by title or window ID) to analyze UI state or extract visual information.

• 2

  Discover and enumerate visible windows on a macOS desktop to find the correct target for visual or interaction tasks.

• 3

  Automate simple keyboard interactions (single keys, modifiers, or typed text) to drive or test macOS applications.

Description

This open-source Model Context Protocol (MCP) server provides window-level screenshot capabilities on macOS, allowing agents to capture screenshots by window title or ID, list and find visible windows, and send keyboard input or type text. It is useful for agents that need visual context from real macOS applications and to drive simple user interactions. The server exposes tools for capturing screenshots in binary or base64, searching windows by title or owner, and simulating key presses and typed text to the active window.