XPath and Selenium Cheat Sheet: Locators, Axes, and Commands

CSS vs XPath (TL;DR)

Before the syntax tables, the one decision every locator starts with: CSS selector or XPath. Default to CSS for speed and readability; switch to XPath when you need to match by text or walk up the DOM.

For the CSS side in depth, see the CSS selectors guide. The rest of this cheat sheet focuses on XPath and Selenium.

XPath Basics: Absolute vs Relative

An absolute XPath starts at the document root with a single slash and names every node down to the target. A relative XPath starts anywhere with a double slash. Absolute paths break the moment the layout shifts, so relative XPath is the default for real automation.

# Absolute XPath - brittle, breaks on any layout change
/html/body/div[2]/form/input[1]

# Relative XPath - resilient, starts with //
//form/input[@id='user-message']

The core relative-XPath syntax, using the Selenium Playground as the target site:

Note that XPath indexes start at 1, not 0, so (//input)[1] is the first match. For the wider set of locator strategies, the Selenium locators guide covers id, name, CSS, and link-text alongside XPath.

XPath Axes and Siblings Navigation

Axes are how XPath moves through the DOM tree relative to a node you can already find. They are the tool for reaching an element that has no useful attributes of its own but sits next to one that does, for example a value cell beside a labeled header.

The most common real-world pattern is following-sibling: find a stable label, then hop to the input or value next to it, for example //label[text()='Email']/following-sibling::input. This survives redesigns that a positional absolute path never would.

XPath Functions and Predicates for Dynamic Elements

Modern apps generate IDs like input_a8f3c that change on every page load, so exact-match locators break constantly. XPath functions solve this by matching on the stable part of an attribute or on visible text.

A robust locator for a dynamic field usually combines a function with a partial value: //input[starts-with(@id,'email_') and @type='email']. For dozens more worked examples, the complete guide to using XPath in Selenium walks through each function on real pages.

Selenium Cheat Sheet: Driver Commands and Locators

XPath finds the element; Selenium WebDriver acts on it. The commands below cover initialization, locating with By.xpath, interaction, and window or frame handling in both Java and Python.

Java (Selenium WebDriver)

import org.openqa.selenium.By;
import org.openqa.selenium.WebDriver;
import org.openqa.selenium.WebElement;
import org.openqa.selenium.chrome.ChromeDriver;

WebDriver driver = new ChromeDriver();
driver.get("https://www.testmuai.com/selenium-playground/simple-form-demo/");

// Locate with XPath and interact
WebElement input = driver.findElement(By.xpath("//input[@id='user-message']"));
input.sendKeys("Hello TestMu AI");
driver.findElement(By.xpath("//button[@id='showInput']")).click();

// Read text back
String output = driver.findElement(By.xpath("//p[@id='message']")).getText();
driver.quit();

Python (Selenium WebDriver)

from selenium import webdriver
from selenium.webdriver.common.by import By

driver = webdriver.Chrome()
driver.get("https://www.testmuai.com/selenium-playground/simple-form-demo/")

driver.find_element(By.XPATH, "//input[@id='user-message']").send_keys("Hello TestMu AI")
driver.find_element(By.XPATH, "//button[@id='showInput']").click()

output = driver.find_element(By.XPATH, "//p[@id='message']").text
driver.quit()

Frequently used WebDriver commands:

These commands run under any Selenium test runner, most commonly TestNG or JUnit in Java and pytest in Python, which handle grouping, assertions, and reporting around the WebDriver calls. To run the suite across real browsers instead of one local machine, point the driver at the TestMu AI cloud grid with a capabilities block; the full flag list is in the Selenium automation capabilities docs, and the language-specific reference lives in the Selenium Python cheat sheet.

Common Selenium Exceptions and Fixes

Most Selenium failures are a handful of exceptions with predictable causes. Knowing the fix turns a red build into a two-line change.

The single highest-value fix is replacing fixed sleeps with explicit waits, so tests proceed the moment an element is ready rather than failing on a slow render or waiting longer than needed.

The Role of AI in Modern QA: Self-Healing Locators

Even a well-written XPath breaks when developers rename an id or restructure the DOM, and locator maintenance is the largest hidden cost in a mature Selenium suite. AI is changing that in three ways:

• Self-healing locators - when a locator fails, the platform records how elements were found in earlier runs and formulates a new locator from those benchmarks, so a renamed class or shifted node no longer breaks the test.
• Automated test generation - AI drafts test cases and assertions from requirements or recorded flows, cutting the time to first coverage.
• Natural-language authoring - tests are written as plain-English instructions instead of hand-coded selectors, opening authoring to the whole team.
• Predictive analytics - AI flags flaky tests and likely failure areas before they block a release.

TestMu AI's Automation Cloud applies self-healing directly: with the auto-healing capability enabled, it detects DOM changes, records the paths of located elements, and reformulates locators when one goes missing, reducing the maintenance the exceptions above create. One honest caveat: healing is heuristic, so for strict regression lanes where any UI change must fail the test, keep stable explicit selectors and leave it off.

For natural-language authoring, KaneAI turns plain-English prompts into executable web, mobile, and API tests and exports them to major frameworks, so the XPath and Selenium syntax in this cheat sheet becomes the layer you review rather than the layer you hand-write.

Wrapping Up

Keep this cheat sheet within reach: default to relative XPath, reach for axes when an element has no attributes of its own, use contains() and starts-with() for dynamic IDs, and replace fixed sleeps with explicit waits to kill most exceptions. Then let AI carry the maintenance, with self-healing locators keeping the suite green as the UI evolves. Start by running your locators across real browsers on the TestMu AI cloud grid, and use the linked XPath and Selenium deep-dive guides whenever you need more than a quick reference.