Advanced GUI Automation: Controlling Windows Apps with pywinauto

We’ve used PyAutoGUI, which is great, but it’s “blind.” It only knows coordinates (e.g., “click at x=500, y=300”). If a window moves, the script breaks. That’s why this Python pywinauto guide will help you move beyond just coordinates and automate GUIs with more reliability.

pywinauto is smarter. It connects to Windows apps and finds elements by their properties (like title="Save" or class_name="Edit"). This is much more reliable.

Note: pywinauto is for Windows only.

Step 1: Installation

pip install pywinauto

Step 2: Finding Your App

First, we need to connect to an application. Let’s automate Notepad.

from pywinauto.application import Application

# 1. Start Notepad
app = Application(backend="uia").start("notepad.exe")

# 2. Connect to the main window
dlg = app.top_window()

Step 3: Controlling the App

Now we can find elements inside the window and control them.

# 1. Find the text editor (its class name is "Edit")
# and type text into it
dlg.child_window(class_name="Edit").set_text("Hello from pywinauto!")

# 2. Select the "File" -> "Save As" menu
dlg.menu_select("File->SaveAs")

# 3. In the "Save As" dialog, find the "File name:" box
save_as_dlg = app.top_window()
save_as_dlg.child_window(title="File name:", class_name="Edit").set_text("test.txt")

# 4. Click the "Save" button
save_as_dlg.child_window(title="Save", class_name="Button").click()

print("File saved!")

This is the most robust way to automate legacy Windows applications that don’t have an API.