Selenium Cheatsheet

Actions API

Use this Selenium reference while you build software engineering projects, review code for technical interview prep, or polish examples for a software engineer resume.

ActionChains Overview

ActionChains queues a series of low-level input actions and dispatches them all with a single .perform() call. Selenium 4 re-implemented the Actions API using the W3C standard input sources.

from selenium.webdriver.common.action_chains import ActionChains
from selenium.webdriver.common.keys import Keys
from selenium.webdriver.common.by import By

ac = ActionChains(driver)

Mouse Actions

Click Variants

el = driver.find_element(By.ID, "target")

ActionChains(driver).click(el).perform()                  # left-click
ActionChains(driver).double_click(el).perform()           # double-click
ActionChains(driver).context_click(el).perform()          # right-click
ActionChains(driver).click_and_hold(el).perform()         # press and hold
ActionChains(driver).release(el).perform()                # release held button

Move to Element

ActionChains(driver).move_to_element(el).perform()                        # hover

# Offset from element center (pixels)
ActionChains(driver).move_to_element_with_offset(el, 10, -5).perform()

# Offset from current mouse position
ActionChains(driver).move_by_offset(100, 50).perform()

Drag and Drop

source = driver.find_element(By.ID, "drag-item")
target = driver.find_element(By.ID, "drop-zone")

# Simple drag-and-drop
ActionChains(driver).drag_and_drop(source, target).perform()

# Drag by offset (pixels from source position)
ActionChains(driver).drag_and_drop_by_offset(source, 200, 0).perform()

# Manual hold-move-release (more reliable in some browsers)
ActionChains(driver)\
    .click_and_hold(source)\
    .pause(0.3)\
    .move_to_element(target)\
    .pause(0.3)\
    .release()\
    .perform()

Scroll (Selenium 4.2+)

from selenium.webdriver.common.action_chains import ActionChains

# Scroll the viewport by delta pixels
ActionChains(driver).scroll_by_amount(0, 500).perform()          # scroll down 500px

# Scroll an element into view
ActionChains(driver).scroll_to_element(el).perform()

# Scroll within a scrollable element
container = driver.find_element(By.ID, "scrollable-div")
ActionChains(driver).scroll_to_element(container).scroll_by_amount(0, 300).perform()

Keyboard Actions

# Key down / key up (hold modifier keys)
ActionChains(driver).key_down(Keys.SHIFT).send_keys("hello").key_up(Keys.SHIFT).perform()
# Sends "HELLO"

# Send keys to focused element
ActionChains(driver).send_keys("some text").perform()

# Send keys to a specific element
ActionChains(driver).send_keys_to_element(el, "typed text").perform()

# Keyboard shortcut
ActionChains(driver)\
    .key_down(Keys.CONTROL)\
    .send_keys("a")\
    .key_up(Keys.CONTROL)\
    .perform()  # Ctrl+A (select all)

Common Key Sequences

# Select all and delete
ActionChains(driver)\
    .click(field)\
    .key_down(Keys.CONTROL).send_keys("a").key_up(Keys.CONTROL)\
    .send_keys(Keys.DELETE)\
    .perform()

# Shift+click (multi-select)
first = driver.find_element(By.CSS_SELECTOR, "li:first-child")
last  = driver.find_element(By.CSS_SELECTOR, "li:last-child")
ActionChains(driver)\
    .click(first)\
    .key_down(Keys.SHIFT)\
    .click(last)\
    .key_up(Keys.SHIFT)\
    .perform()

# Ctrl+click (toggle selection)
items = driver.find_elements(By.CSS_SELECTOR, "li.item")
ActionChains(driver)\
    .click(items[0])\
    .key_down(Keys.CONTROL)\
    .click(items[2])\
    .click(items[4])\
    .key_up(Keys.CONTROL)\
    .perform()

Chaining Actions

# All steps queued and executed together
ActionChains(driver)\
    .move_to_element(menu)\
    .pause(0.5)\
    .click(submenu_trigger)\
    .pause(0.3)\
    .click(menu_item)\
    .perform()

W3C Actions API (Selenium 4 — Low Level)

from selenium.webdriver.common.actions.wheel_input import ScrollOrigin

# Scroll from a specific element origin
scroll_origin = ScrollOrigin.from_element(el)
ActionChains(driver).scroll_from_origin(scroll_origin, 0, 200).perform()

# Scroll from element with offset
scroll_origin = ScrollOrigin.from_element(el, 0, -50)
ActionChains(driver).scroll_from_origin(scroll_origin, 0, 100).perform()

# Viewport-origin scroll
scroll_origin = ScrollOrigin.from_viewport(10, 10)
ActionChains(driver).scroll_from_origin(scroll_origin, 0, 400).perform()

Pause

# Insert a pause (seconds) between chained actions
ActionChains(driver)\
    .click(el)\
    .pause(1.0)\
    .send_keys("text")\
    .perform()

Reset Actions

ac = ActionChains(driver)
ac.click(el)
ac.reset_actions()   # clear the queued actions without performing

Gotchas

drag_and_drop can fail on HTML5 drag-and-drop implementations because they rely on synthetic drag events that browsers may ignore. Use a JavaScript-based drag helper or the click_and_hold → move_to_element → release approach.

key_down must always be paired with key_up — an un-released modifier key persists in the browser session and breaks subsequent actions.

Mouse offsets are from the element's center when using move_to_element_with_offset. Negative values move up/left.

perform() is not cumulative — calling it does not clear the internal queue automatically in older versions. Call reset_actions() or create a new ActionChains instance for each sequence.