Using pywinauto for Windows GUI Automation with Python

Today we share a Python library for automating mouse and keyboard actions on Windows GUI applications: pywinauto .

1. Installation

Install the library via pip:

pip install pywinauto

Official documentation: https://pywinauto.readthedocs.io/en/latest/

2. Application

To control a Windows program, create an Application object. The optional backend parameter can be win32 (default) or uia . win32 works with MFC, VB6, VCL, simple WinForms, and legacy apps; uia works with WinForms, WPF, Store apps, Qt5, browsers, etc.

If you are unsure which backend to use, inspect the target with Inspect (for uia ) or Spy++ (for win32 ) and choose the one that shows more information.

Typical methods of the Application object are shown in the following diagram:

Example: start the WeChat application by connecting to its process ID (PID) shown in Task Manager.

3. WindowSpecification

Each window is represented by a WindowSpecification object, obtained via Application.window() using parameters such as title , classname , or best_match . Note that the name shown in Inspect corresponds to the title argument.

Common methods of WindowSpecification are illustrated in the following image:

4. Element Controls

Windows contain many control types (Button, Edit, TreeView, CheckBox, Dialog, Toolbar, StatusBar, ListBox, Pane, Menu, Header, etc.). Each control has an element_info attribute that returns an ElementInfo derived object (e.g., UIAElementInfo or HwndElementInfo ) with useful properties and methods.

Common wrapper methods are listed in the following tables:

Two particularly useful methods are click_input() (mouse click) and type_keys() (keyboard input).

Mouse Operations

The mouse module provides the core private method _perform_click_input() , wrapped by higher‑level functions such as click_input() . Important parameters include:

button : default "left"; options are left, right, middle, move, wheel, x.

coords : a tuple (x, y) specifying screen coordinates; default (0, 0).

wheel_dist : scroll distance; positive for up, negative for down.

Example screenshot of mouse click usage:

Keyboard Operations

The keyboard module’s main function is send_keys() . The first argument keys specifies the key sequence. Optional boolean flags (e.g., with_spaces , with_tabs , turn_off_numlock ) control behavior.

Keys are expressed inside curly braces, e.g., {ENTER} . By default, keys are sent as virtual‑key codes prefixed with VK_ . To send literal characters, set vk_packet=False .

Modifiers can be written with down and up inside braces, or using shorthand symbols: + for Shift, ^ for Control, % for Alt.

Example of modifier usage:

For a full list of supported keys, refer to the official documentation: pywinauto.keyboard .

In summary, pywinauto offers a comprehensive set of APIs to launch applications, locate windows and controls, and simulate mouse and keyboard actions on Windows, making it a powerful tool for GUI testing and automation.