TestMu AI brings natural-language mobile testing to Kane CLI

TestMu AI, formerly known as LambdaTest, has announced mobile automation support in Kane CLI, its agentic test-execution tool. The update extends the same vision-grounded reasoning engine the company uses for web testing to iOS Simulators and Android Emulators, allowing engineers to describe an expected user outcome in plain English and have the agent drive the device, verify the rendered screen, and return a structured verdict.

The company says it deliberately avoided wrapping Appium, the dominant open-source mobile automation framework, and instead extended Kane's core perception graph directly to native mobile viewports. At each test step, the agent captures a screenshot and reads the native accessibility document, serialising each on-screen element with its affordance, geometry, state and identity. This approach covers native UI frameworks including UIKit, SwiftUI, Jetpack Compose, Flutter and React Native without separate configuration for each.

The technical case

One of the more persistent pain points in mobile test automation is hybrid-app context switching: when a screen mixes native views with embedded WebViews, tools built on Appium typically require an explicit command to switch contexts, which is a common source of test flakiness. TestMu AI says Kane CLI folds WebView elements into the same perception list as native elements, removing the explicit switch entirely. For non-standard components such as custom canvas elements or Flutter controls, the agent falls back to visual reasoning and coordinate-grounding to keep the flow moving.

Each test run produces what the company calls an "evidence pack": step-annotated screenshots, native device logs and step-attributed network traffic bundled into a portable artefact. Mudit Singh, co-founder and head of growth, argued that this changes the nature of the output. "What you get back isn't a claim about what the agent says it did. It's a verified verdict, with the screenshots, device logs, and network traffic to back it up."

Market context

Mobile UI testing has been a recognised productivity drain for more than a decade. The standard toolchain, built around Appium, Xcode Instruments and Android Debug Bridge, requires substantial setup time and generates brittle locator-based tests that break whenever UI element identifiers change across builds. A number of well-funded AI-native testing startups are pursuing similar approaches to test-maintenance reduction, including vision-based and large-language-model-assisted test generation, putting pressure on incumbents in the quality-engineering category.

The broader shift toward agentic AI in developer tooling is gathering pace, with vendors across CI/CD, observability and testing repositioning their products around autonomous agents rather than script-based automation. Enterprise software buyers in this space are increasingly scrutinising how agents handle flaky test scenarios, how auditable their verdicts are, and whether the evidence trail is sufficient for regulated-industry compliance teams.

TestMu AI's rebrand from LambdaTest, a name with significant recognition in the browser testing market, signals an intent to lead on the agentic angle rather than compete primarily on device-cloud scale. The company is headquartered in Noida, India, with a San Francisco presence, and has not disclosed revenue, customer counts or funding details in this release. Mobile Automation in Kane CLI is available immediately; support is currently scoped to simulators and emulators rather than physical device farms, which will be a threshold question for enterprise buyers with real-device testing requirements.