TestMu AI

TestMu AI (formerly LambdaTest) runs Mobilewright tests on real Android and iOS devices through the @testmuai/mobilewright driver, maintained by TestMu AI. Each test shows up on the App Automation dashboard with its video, device logs and network logs, named after the test and marked passed or failed, and no TestMu AI-specific code is needed in your tests.

Requires mobilewright 0.0.56 or later and Node.js ≥ 22.12. Real devices only, emulators and simulators are not supported. Mobilewright support on TestMu AI is in beta.

Setup

1. Install the driver

npm install --save-dev @testmuai/mobilewright

2. Set your credentials

Find your username and access key under Account Settings > Password & Security and export them:

export TESTMU_USERNAME=...
export TESTMU_ACCESS_KEY=...

The LT_USERNAME and LT_ACCESS_KEY names also work. The driver takes username and accessKey options directly too, but keep credentials out of the repository.

3. Upload your app

Every session starts with the app under test installed. Upload it once and keep the lt:// ID it returns:

curl -u "$TESTMU_USERNAME:$TESTMU_ACCESS_KEY" \
  -X POST https://manual-api.lambdatest.com/app/upload/realDevice \
  -F "appFile=@app.apk"

The app_url in the response is the ID to pass to the driver. You can also pass a local .apk, .aab or .ipa path, which is uploaded once per run and shared across workers, or a public https:// URL.

4. Set the driver

// mobilewright.config.ts
import { defineConfig } from 'mobilewright';
import { testMuDriver } from '@testmuai/mobilewright';

export default defineConfig({
  testDir: './tests',
  timeout: 180_000,
  use: { actionTimeout: 20_000 },
  expect: { timeout: 20_000 },
  driver: testMuDriver({
    apps: {
      android: 'lt://APP_ANDROID_ID',
      ios: 'lt://APP_IOS_ID',
    },
    build: 'my-first-build',
  }),
  projects: [
    {
      name: 'android',
      use: {
        platform: 'android',
        deviceName: /Pixel/,
        osVersion: '>=13',
        bundleId: 'com.lambdatest.proverbial',
      },
    },
    {
      name: 'ios',
      use: {
        platform: 'ios',
        deviceName: /iPhone/,
        osVersion: '>=17',
        bundleId: 'com.lambdatest.proverbial',
      },
    },
  ],
});

Device selection uses the standard config fields: platform, deviceName (regex) and osVersion (exact, or a range like ">=17 <19", rounded to whole major versions). Broader patterns get a device faster. Set bundleId so Mobilewright can relaunch the app before each test.

The app under test goes in the driver's app or apps, not in installApps. Use installApps only for helper apps.

Mobilewright's 5-second default timeouts are meant for local devices. Raise actionTimeout and expect for cloud devices, as in the example above.

Driver options

All options are optional, but every session needs an app.

Option	Purpose
username, accessKey	Credentials; default to the environment variables above
app	lt:// ID, local file or https:// URL; defaults to TESTMU_APP
apps	Per-platform apps keyed android, ios, android-real, ios-real; overrides app
build, project, name, tags	How sessions are grouped and named on the dashboard; build defaults to TESTMU_BUILD
region	US, EU or AP; defaults to the nearest data center
video, deviceLog, networkLog	What TestMu AI records per session
tunnel, tunnelName	Route device traffic through a running tunnel
geoLocation, timezone	Two-letter country code for IP geolocation, device timezone
autoGrantPermissions	Grant Android runtime permissions on install
autoAcceptAlerts, autoDismissAlerts	Handle iOS system alerts
disableAnimation	Turn off system animations (use.animations has no effect here)
visibility	'native' (default) or 'bounds', which matches local runs
allocationTimeout	How long to wait for a device, including queue time (default 900s)
capabilities	Any other App Automation capability, such as networkProfile; wins over named options

The references page lists every option and its default.

Running tests

Each worker uses one device. Set workers to your plan's parallel session limit, extra workers wait for a free device. Run Android and iOS in separate processes:

npx mobilewright test --project android
npx mobilewright test --project ios

DEBUG=testmu:* npx mobilewright test   # driver debug logging

The debug log shows device allocation, every device command and app uploads. Include it and the session ID when you contact TestMu AI support.

Tunnels

Use the TestMu AI tunnel when your app talks to a server on your machine or on a private network. Start the tunnel binary first, then point the driver at it by name:

./LT --user "$TESTMU_USERNAME" --key "$TESTMU_ACCESS_KEY" --tunnelName my-tunnel

driver: testMuDriver({
  app: 'lt://APP_ID',
  tunnel: true,
  tunnelName: 'my-tunnel',
}),

Known differences from local runs

• On iOS, an element covered by another one can report as not visible. Set visibility: 'bounds' to match local runs.
• getByType() matches raw native types such as XCUIElementTypeSwitch. Prefer getByRole().
• Web views are visible only in debug builds, or on iOS 16.4+ with isInspectable = true.
• device.listApps() returns only apps launched in the current session.

Switching between a local device and TestMu AI

Keep one config and pick the driver by environment: local (mobilecli) by default, TestMu AI when credentials are present.

import { defineConfig, type MobilewrightConfig } from 'mobilewright';
import { testMuDriver } from '@testmuai/mobilewright';

const config: MobilewrightConfig = { /* ...your existing config... */ };

if (process.env.TESTMU_USERNAME) {
  config.driver = testMuDriver({ app: process.env.TESTMU_APP });
}

export default defineConfig(config);

npx mobilewright test                                   # local device
TESTMU_USERNAME=... TESTMU_ACCESS_KEY=... TESTMU_APP=lt://... \
  npx mobilewright test                                 # the same suite on TestMu AI

Learn more

Setup guides, troubleshooting and a complete example project are in the TestMu AI Mobilewright docs and the sample repository.