pyautogui: Automate Mouse and Keyboard Actions in Python

Overview

PyAutoGUI is a Python library for automating desktop interactions. It hides platform-specific APIs behind a shared interface for mouse movement and clicks, keyboard input, message boxes, and screenshot-based image matching.

It is aimed at people automating tasks in graphical applications rather than interacting through an application-specific API. The project supports Windows, macOS, and Linux, with platform-dependent setup and limitations such as primary-monitor support.

Key Features

• Move the mouse to screen coordinates or relative to its current position, and perform clicks and double-clicks.
• Type text, press keys, and send keyboard shortcuts to the currently focused window.
• Add durations and easing functions to mouse movements.
• Display alert, confirmation, prompt, and password dialogs.
• Capture screenshots and save them as image files.
• Locate an image on screen and return its position for follow-up actions such as clicking.

Use Cases

• Automate repetitive desktop steps in a graphical application when it has no suitable API for the task.
• Create small personal workflows that type text, navigate controls, or use keyboard shortcuts.
• Build image-guided scripts that find a visible button or control before interacting with it.
• Capture screenshots as part of a desktop workflow or a lightweight visual check.

Getting Started

Install from PyPI:

pip install pyautogui

See the README and documentation for platform setup and API details.

Alternatives

• open-codex-computer-use: It exposes desktop automation through an MCP service and accessibility controls, rather than a direct Python mouse-and-keyboard API.
• Open-Interface: It turns natural-language requests into LLM-guided mouse and keyboard actions, rather than providing a general-purpose Python control library.
• UFO: It focuses on Windows desktop agents and coordinated, multi-device workflows, rather than direct cross-platform desktop control from Python.
• Agent-S: It uses screenshots and an AI agent to perform natural-language desktop tasks, rather than offering direct mouse and keyboard primitives.

Considerations

The project is marked inactive, with no commits or releases in over a year. Although it has dozens of contributors, contributions are highly concentrated in one person; recent issue handling and pull request activity are also sparse. Teams adopting it should plan to test their workflows on their target operating systems and be prepared to maintain or work around issues themselves.

PyAutoGUI currently works only on the primary monitor, and its README cautions that second-monitor mouse functions may or may not work depending on the operating system and version. Setup differs by platform: macOS requires PyObjC modules, Linux requires Xlib, and image features use Pillow. The README also notes that Linux may need additional libraries for Pillow image support.