How does it address common automation challenges like focus-stealing and input methods (IME)?

It prevents focus-stealing by automatically refocusing the target application before executing actions via an `app` parameter. For IME, especially CJK languages, it offers a 'direct' typing mode via AppleScript to completely bypass input methods for consistent text entry.

Is macOS Desktop Control compatible with popular AI development environments or clients?

Yes, it works with any MCP-compatible client that uses stdio transport. This includes popular tools like Claude Code, Claude Desktop, OpenAI Codex CLI, Gemini CLI, Cursor, and VS Code Copilot.

What kind of deep macOS integrations does this tool offer?

It supports running macOS Shortcuts workflows and provides comprehensive keyboard control using AppleScript key codes for any key and modifier combination. It also leverages AppleScript for reliable window management.

What is macOS Desktop Control?

It's an MCP server designed for native macOS desktop automation, providing precise control over screenshots, mouse, keyboard, and window management directly on your actual Mac desktop. It's ideal for AI agents and scripts.

macOS Desktop Control: Native Automation for AI Agents

This server offers robust native macOS desktop automation, allowing seamless control over screen actions, mouse movements, keyboard inputs, and window management without needing Docker or a virtual display. It empowers AI clients like Claude Code, OpenAI Codex CLI, Gemini CLI, and Cursor to interact directly with your actual macOS environment. Key features address common challenges such as preventing focus-stealing during AI tool approvals and bypassing input methods (IME) for reliable text input, making it an indispensable tool for advanced macOS automation.

Características Principales

01Native macOS Desktop Automation: Full control over screenshots, mouse (click, move, drag, scroll), keyboard (type text, key combos), and window management (list, focus, open apps).

02Focus-Stealing Prevention: Automatically refocuses the target application before executing actions, ensuring commands land in the correct window.

03Input Method (IME) Bypass: Offers a "direct" text typing mode via AppleScript to completely bypass input methods for consistent text entry, especially useful for CJK languages.

04Deep macOS Integration: Supports running macOS Shortcuts workflows and provides comprehensive keyboard control using AppleScript key codes for any key + modifier combination.

05Lightweight Dependencies: Relies primarily on `cliclick` and built-in macOS tools like `osascript` and `screencapture`, avoiding heavy runtime dependencies.

060 GitHub stars

Casos de Uso

01Empowering AI agents (e.g., Claude Code, OpenAI Codex CLI) to interact with and control the macOS desktop environment.

02Automating repetitive system tasks and application workflows directly on a native macOS installation.

03Developing sophisticated desktop automation scripts that require precise control over application focus and input methods.