Small Windows automation toolkit for screen, mouse, OCR, image matching, colors, windows, and shell commands

simpleautogui is a Python library for pragmatic Windows GUI automation. It wraps common low-level tools like PyAutoGUI, PyWin32, Pillow, OpenCV-compatible image search, and Tesseract OCR behind a small object-oriented API.

The library is useful when a workflow is easier to automate through the real desktop than through an official API: click a known screen point, wait for an image, read text from a region, find a color, move windows, or run a Windows command.

Navigation

• Installation
• Core concepts
• Point
• Region
• OCR
• Image matching
• Color matching
• Macros and real usage
• Windows
• Window grids
• Monitors
• Console commands
• Development

Installation

pip install simpleautogui

Requirements:

• Python 3.11+
• Windows
• Tesseract OCR installed separately if you use OCR methods

For development:

pip install "simpleautogui[dev]"

Core concepts

The public API is intentionally small:

• Point represents one screen coordinate.
• Region represents a rectangular screen area.
• Window wraps a native Windows window handle.
• WindowsGrid arranges several windows inside a region.
• Monitor describes available displays.
• AbstractMacro and MacroRunner run reusable automation scripts by hotkeys.
• cmd and powershell run Windows shell commands.

from simpleautogui import AbstractMacro, MacroRunner, Point, Region, Window, WindowsGrid, Monitor, cmd, powershell

Coordinates follow the usual screen convention: x grows left to right, y grows top to bottom. Region(x, y, w, h) stores width and height, not right/bottom coordinates.

Point

Use Point when you already know the coordinate you want to click, move to, or drag from.

from simpleautogui import Point

start = Point(100, 100)
target = Point(500, 500)

start.click()
start.move_in(duration=0.2)
start.drag_to(target, duration=0.5)
start.drag_rel(200, 0, duration=0.3)

print(start.to_tuple())  # (100, 100)

If x or y is omitted, the missing coordinate is taken from the current cursor position. Zero is a valid coordinate:

Point(0, 100)
Point(100, 0)
Point(None, 100)

Read the pixel color under a point:

color = Point(100, 100).color
print(color)  # (r, g, b)

Wait for a mouse or keyboard input and return the cursor position:

point = Point.input(button="right", timeout=10)
print(point)

Region

Use Region when an action belongs to a rectangular screen area.

from simpleautogui import Region

region = Region(100, 200, 300, 400)
fullscreen = Region()

region.click()
region.click(center=False)
region.move_in(o_x=10, o_y=5)
region.show()

print(region.to_tuple())  # (100, 200, 300, 400)

Drag from a region:

from simpleautogui import Point, Region

panel = Region(100, 100, 400, 300)
panel.drag_to(Point(800, 300), center=True, duration=0.4)
panel.drag_rel(200, 0, center=False, duration=0.4)

Filter close duplicate regions:

unique_regions = Region.remove_proximity(regions, proximity_threshold_px=10)

OCR

OCR uses pytesseract. Install Tesseract OCR and make sure tesseract.exe is available in PATH, or configure pytesseract in your application.

Read text from a region:

from simpleautogui import Region

text = Region(100, 100, 600, 200).text(
    lang="eng+rus",
    resize=2,
    contrast=1.5,
    sharpen=True,
)

if "Ready" in text:
    print("Application is ready")

Find a word or a phrase and get matching regions:

from simpleautogui import Region

matches = Region().find_text(
    text="Export complete",
    lang="eng",
    resize=2,
    min_confidence=80,
    case_sensitive=False,
)

for match in matches:
    match.click()

resize improves OCR quality on small UI text. Returned coordinates are scaled back to the original screen coordinates.

Image matching

Use image matching when the UI element is easier to identify by screenshot than by text.

from simpleautogui import Region

button = Region().wait_image(
    paths="assets/export_button.png",
    timeout=10,
    confidence=0.9,
    check_interval=0.1,
)

if button:
    button.click()

Wait for one of several images:

result = Region().wait_image(
    paths=("assets/ok.png", "assets/continue.png"),
    timeout=15,
)

Find multiple matches:

icons = Region().wait_images(
    paths="assets/item.png",
    timeout=10,
    confidence=0.9,
    proximity_threshold_px=2,
    min_matches=1,
)

for icon in icons:
    icon.click()

Color matching

Color matching is useful for simple UI state checks: active indicator, progress color, badge color, selected state.

from simpleautogui import Region

point = Region().wait_color("red", timeout=10, confidence=0.95)

if point:
    point.click()

Supported color formats:

• "#ff0000"
• "#f00"
• "rgb(255, 0, 0)"
• "red"
• (255, 0, 0)

Find several colors:

points = Region(0, 0, 800, 600).wait_colors(
    color=("#00ff00", "rgb(255, 0, 0)", "blue"),
    timeout=10,
    confidence=0.9,
    check_interval=0.1,
    proximity_threshold_px=2,
    min_matches=0,
)

if points:
    print(points[0])

confidence=1 means exact match. Lower confidence increases the RGB tolerance.

Macros and real usage

Use AbstractMacro when you want to start and stop an automation script from any screen with keyboard shortcuts. The macro itself runs in a worker thread. Stop is cooperative: the runner requests stop, and the macro exits at the next context.check_stop(), context.sleep(...), context.wait_image(...), or context.wait_color(...) call.

from simpleautogui import AbstractMacro, MacroContext, MacroRunner, Region


class ClickImagesMacro(AbstractMacro):
    def run(self, context: MacroContext) -> None:
        screen = Region()

        while context.is_running:
            button = context.wait_image(
                region=screen,
                paths=("assets/next.png", "assets/continue.png"),
                timeout=1,
                confidence=0.9,
            )

            context.check_stop()

            if button:
                button.click()
                context.sleep(0.2)


if __name__ == "__main__":
    runner = MacroRunner(
        ClickImagesMacro(),
        start_hotkey="ctrl+alt+s",
        stop_hotkey="ctrl+alt+q",
    )

    print("Ctrl+Alt+S - start, Ctrl+Alt+Q - stop, Esc - exit")
    runner.listen(exit_hotkey="esc")

Typical workflow:

• run the Python script once;
• press Ctrl+Alt+S to start the macro;
• press Ctrl+Alt+Q to request stop;
• press Esc to unbind hotkeys and exit the listener.

You can also use one toggle shortcut:

runner = MacroRunner(ClickImagesMacro(), toggle_hotkey="ctrl+alt+m")
runner.listen(exit_hotkey="esc")

Lifecycle hooks are optional:

class MyMacro(AbstractMacro):
    def on_start(self, context: MacroContext) -> None:
        print("started")

    def run(self, context: MacroContext) -> None:
        ...

    def on_stop(self, context: MacroContext) -> None:
        print("stopped")

If your macro uses long loops or long waits, prefer context.sleep, context.wait_image, context.wait_images, context.wait_color, and context.wait_colors over direct long blocking calls. That keeps hotkey stop responsive.

Windows

Use Window to find and control native Windows windows.

from simpleautogui import Window

window = Window(title="Notepad")
print(window.title)
print(window.region)

window.raise_it()
window.set_geometry(x=0, y=0, w=900, h=700)
window.maximize()
window.restore()
window.minimize()
window.close()

Find all visible windows or filter by title:

from simpleautogui import Window

for window in Window.all():
    print(window.title)

notepads = Window.by_title("notepad", case_sensitive=False)

Window grids

WindowsGrid arranges windows inside a target region.

from simpleautogui import Region, Window, WindowsGrid

windows = Window.by_title("notepad")

grid = WindowsGrid(
    windows=windows[:4],
    rows=2,
    cols=2,
    region=Region(0, 0, 1600, 900),
)

grid.arrange()

append, prepend, and insert update the window list and arrange the grid again:

grid.append(Window(title="Calculator"))
grid.prepend(Window(title="Explorer"))
grid.insert(1, Window(title="Terminal"))

Monitors

Monitor exposes full and work regions for each display.

from simpleautogui import Monitor

for monitor in Monitor.all():
    print(monitor.name)
    print(monitor.fregion)  # full monitor region
    print(monitor.wregion)  # work area without taskbar
    print(monitor.flags)
    print(monitor.device)

Console commands

Run Windows shell commands and get stdout as a string.

from simpleautogui import cmd, powershell
from simpleautogui.win.console.exceptions.base import CommandExecutionError

try:
    print(cmd("dir", timeout=10))
    print(powershell("Get-ChildItem", timeout=10))
except CommandExecutionError as exc:
    print(exc)

For safer cmd calls, pass a list instead of a shell string:

output = cmd(["cmd.exe", "/c", "dir"], timeout=10)

Development

Install development dependencies:

pip install -e ".[dev]"

Run checks:

python -m ruff check .
python -m pytest -q
python -m build
python -m twine check dist/*

Release workflow is described in RELEASE_GUIDE.md.