Broken Locators

The button is still there. Your locator just cannot find it anymore.

When the application works but the test fails because an ID, XPath, CSS, role or accessibility attribute changed, self-healing can reduce maintenance noise.

Element not foundLocator DriftAuto Recovery
TestZombie AI

Common causes of broken locators

Refactoring, new components, dynamic IDs, changed labels or a different DOM/view hierarchy can turn valid user flows red.

01

Web locators

Stabilize Selenium and Playwright when IDs, CSS, XPath, roles, labels and text change.

02

Mobile locators

Support Appium when accessibility IDs, resource IDs and mobile UI hierarchies change.

03

Do not replace blindly

Candidates are evaluated from multiple signals; if confidence is too low, the failure remains.

Workflow

Turn “element not found” into controlled healing

1

Detect the failure

The test reports that the original locator no longer returns the expected element.

2

Reconstruct the target

TestZombie uses existing context to evaluate potential target elements.

3

Document healing

A robust replacement can be used and stored for review or source-code update.

FAQ

Broken locator FAQ

Is a broken locator automatically a flaky test?

Not necessarily. A locator can become deterministically broken after a UI change, while self-healing still addresses the same maintenance problem.

Which frameworks does TestZombie cover?

The public product experience presents self-healing for Selenium, Playwright and Appium.

What happens with an uncertain candidate?

If no sufficiently confident replacement exists, the test should fail rather than use the wrong element.

TestZombie AI

Your tests should survive change.

Add healing, analysis and actionable fix recommendations to existing Selenium, Playwright and Appium tests.