Real Device Cloud Testing: How to Test Mobile Apps on Devices Virtually

Originally published on the Momentic blog.

It’s important to understand what real device cloud testing is and how it works. It’s also important to understand that it’s not the only way to test your app across a range of devices. In fact, in some instances it may well hold you back.

This article guides you through the basics of real device testing and how it works, then discusses its pros and cons compared to alternatives. For teams that need toscale efficiently or maintain rapid release cycles, conducting most testing on emulators or simulators may be a more practical option.

Summary: Real Device Cloud Testing and Its Alternatives

Here’s a quick summary of the key takeaways from this blog. We’ll cover each of these in more detail below.

Topic	Key takeaway
Real device testing	Tests app functionality and performance on physical devices.
Real device clouds	Provide remote access to physical devices without maintaining your own device farm.
Virtual testing	Emulators and simulators provide virtual environments for testing across devices and OS versions.
Real vs. virtual	Real devices capture hardware-specific behaviour; virtual devices offer scalable testing without physical hardware.
Cost & scale	Real device testing can become costly and introduce device availability constraints as testing scales.
Emulators & simulators	Useful for broad, routine and automated testing across a range of virtual environments.
Where real devices matter	Physical devices remain useful for validating sensors, OEM-specific behaviour and other hardware-dependent issues.
Modern QA	Virtual testing can support high-volume, AI-driven testing and rapid release cycles.

Real device testing uses physical devices to verify whether your app works as intended. To test on real devices, you can either run your own device farm if you have the expertise and resources, or use a third-party real device cloud.

Real device testing provides the most accurate picture of app performance, but may be slower and expensive to scale. For teams that this won’t work for, there are alternatives.Emulators and simulators allow you to test most aspects of your app virtually; although there are OEM-specific behaviors you can’t verify, today’s options allow you to test to a high degree of accuracy without the need for extensive real device testing.

What is Real Device Testing?

Real device testing is the process of verifying your apps’ functionality and performance on real devices.

Thanks to a diversification of available combinations of devices, platforms, operating systems, screen sizes, and screen resolutions, your user base is likely to access your app on a broad range of devices. These will affect how your UI displays and how certain features work.

To avoid alienating chunks of your workforce, you need to be sure that your app works equally well across likely device/platform/OS/screen combos. Some teams choose to test this on real devices because some issues are difficult to reproduce elsewhere, including:

• Touch and physical interaction
• Network conditions
• Hardware-dependent bugs
• Real-world usability
• Some performance issues (frame drops, slow startup, memory pressure, battery drain)

What is a Real Device Cloud?

A real device cloud is a service that lets you test your app on physical devices hosted remotely in a data centre.

So, rather than maintaining a device farm as part of your own infrastructure, you can connect to a service that offers thousands of them through a web interface or testing API. This avoids the time/resource drain of device farm hardware maintenance.

For example, using a real device cloud would allow you totest a native Android app across Samsung, Google, and other Android phones across different Android versions, without having to actually own or maintain the devices yourself. You could similarlytest iOS apps across different iPhones and iOS versions.

There is a trade-off, however. Third-party real device clouds are paid-for services, and costs can increase significantly as you scale. You may also experience queues for the most in-demand device and platform combinations, slowing down your release cycles.

How Does Real Device Cloud Testing Work?

Here’s a quick, five-step overview of how you carry out real device cloud testing. The slowdowns mentioned above tend to happen between steps two and three; if other device farm users are already testing on the devices you need, you’ll have to wait your turn.

1. You upload your app
2. The cloud gives you access to real devices
3. You install/run the app on those devices
4. You interact with the app remotely or run automated tests
5. The service captures results such as screenshots, videos, logs, crashes, and performance data

When you need a real device, and when you don't

Test needs	Real device	Simulator or emulator
UI flows, navigation, forms, login	Not needed	Yes
Layout across screen sizes and OS versions	Not needed	Yes
Deep links, push notification handling (simulated)	Not needed	Yes, on recent iOS simulators and Android emulators
Camera, NFC, Bluetooth, biometrics with real hardware	Yes	No, or only mocked
Carrier networks, SIM, SMS	Yes	No
Battery, thermal throttling, real performance numbers	Yes	No
OEM-specific Android skins and bugs (Samsung, Xiaomi)	Yes	Partly, with vendor images

Most teams run functional regression tests on simulators and emulators on every pull request, because they start faster and run in parallel cheaply. They keep a smaller real-device pass for hardware features and before major releases.

For how the two virtual options differ, seeemulators vs simulators in mobile testing.

Real device cloud providers compared

Provider	Devices	Automation frameworks	Pricing model
BrowserStack App Automate	Real iOS and Android devices	Appium, Espresso, XCUITest, Maestro, Flutter and others	Subscription, priced by parallel sessions
Sauce Labs	Real devices plus emulators and simulators	Appium, Espresso, XCUITest	Subscription, priced by concurrency
LambdaTest	Real devices plus emulators and simulators	Appium, Espresso, XCUITest	Subscription, priced by parallel sessions
AWS Device Farm	Real devices	Appium, Espresso, XCUITest	Pay per device minute after a one-time free trial, or a flat monthly rate per unmetered device slot
Firebase Test Lab	Real and virtual Android devices, real iOS devices	Espresso and other instrumentation tests, Robo, XCTest/XCUITest, Game Loop	Free daily quota, then billed per device hour, per minute

Firebase now labels Test Lab as deprecated in its product navigation, so we don't recommend it for new setups. For more options, seeBrowserStack alternatives andSauce Labs alternatives.

Are There Alternatives to Real Device Testing?

Yes. Instead of using a real device cloud for testing, you can test your apps virtually using emulators and simulators. These aim to recreate real device environments as closely as possible while providing a faster, lower-cost alternative to real device clouds that’s easier to scale.

Some open-source emulators and simulators are free to use. They are also available as native functionality on somemobile testing platforms, such asMomentic.

Why Use Virtual Device Testing Instead of Real Device Testing?

It’s not simply a case of either/or. Many teams use emulators and simulators for bulk routine testing tasks, then use a limited amount of real device testing for final validation and OEM-specific behavior.

Why? Extensive real device cloud testing isn’t always practical because it incurs high costs as you scale, and device queues slow your release cycles down. In the current landscape, where engineering teams are under pressure to accelerate release cycles while validatingunprecedented amounts of AI-generated code, it’s not a strong fit for many teams.

And as technology evolves, virtual device testing can replicate real device environments with an increasing degree of accuracy. For example, AI can synthesize realistic sensor inputs, such as GPS traces, motion patterns, and network fluctuations, then feed these into emulators to produce test conditions that mimic real-world usage.

This makes them a natural companion for autonomous, AI-driven QA processes because:

• They can provide the realistic execution environment needed for AI tools to surface meaningful issues
• Agentic QA workflows help teams process the volume of results from emulator-based testing (particularly useful for testing reams of AI-generated product code)
• AI models can prioritize emulator test runs based on code changes, historical failures, and usage data

Emulators and Simulators: How Do They Work?

The terms ‘emulator’ and ‘simulator’ aren’t interchangeable (though many non-engineers use them that way).

An emulator recreates the entire hardware and software environment of a real device as accurately as possible. It mimics the device’s CPU architecture, peripherals, firmware, and often the operating system behavior.

A simulator reproduces the behavior of a device without modelling hardware specifics. Typically, it’ll run on a host machine’s CPU instructions to simulate system APIs and UI components rather than running the real device OS image.

How to Run Virtual Testing On Emulators and Simulators

1. Choose a virtual device
2. Launch the emulator or simulator on your machine or in a hosted environment
3. Install your app
4. Run tests (manual or automated)
5. Capture results (automated testing platforms can collect screenshots, videos, logs, crashes, and test results for analysis)

The Pros and Cons of Real Device Testing vs Testing on Emulators and Simulators

	Real devices	Emulators	Simulators
Hardware	Runs on physical device hardware	Recreates device hardware and software virtually	Simulates device software and behaviour
Real-world behaviour	Uses actual sensors, peripherals and network conditions	Can simulate sensors, peripherals and network conditions	Simulates software-level behaviour
Performance	Measures performance on physical hardware	Measures performance within a virtualised environment	Measures software behaviour on the host machine
Scale	Requires access to individual physical devices	Creates multiple virtual device environments	Creates multiple simulated environments
Cost	Requires physical devices or access to a device cloud	Can run locally or in the cloud	Can run locally or in the cloud
Typical use	Hardware-specific and final validation	Broad environment and automated testing	UI and software-behaviour testing

What real device testing costs

Three costs grow with the suite: parallel sessions (most clouds price by how many tests can run at once), session time (real devices are slower to boot and reset than emulators), and queue time when the devices you want are busy. A suite that runs in 10 minutes on 20 parallel emulators can take much longer on 5 real devices. Price the parallelism you need, not the device catalogue.

For Android tooling to run on emulators and devices, see thebest Android testing tools.

Virtual Device Testing: How Momentic Helps

Momentic runs mobile tests on iOS simulators and Android emulators, locally or hosted. Tests are YAML files with natural-language steps, run with npx momentic-mobile run, so the same suite runs on a laptop and in CI. Keep a real-device pass for camera, biometrics and carrier features.

Deploy Multiple Times a Day While Minimizing Defects

“We’ve been adamant about leaning on the best tools available to support our team as we scale. That’s why we use Momentic for testing.”

Alex CuiCo-founder and CTO, GPTZero

AI text identification platform GPTZero found thatCypress could not keep up with their demands as their product scaled. Instead of upping the engineering hours spent on QA, they chose to automate end-to-end tests with Momentic’s agentic testing platform.

This resulted in an 89% decrease in defect escape rate, alongside 80% faster release cycles and an acceleration of 75% in test creation times.

Get in touch to see what Momentic could do for your team.

Frequently asked questions

What is real device cloud testing?

Real device cloud testing lets teams remotely test apps on physical devices hosted by a third-party provider.

What is a real device cloud?

A real device cloud is a remotely hosted collection of physical smartphones and tablets used for app testing.

Is real device testing better than emulators?

Real devices are better for hardware-specific validation, while emulators are faster and easier to scale for routine testing.

What are the disadvantages of real device cloud testing?

It can become expensive at scale and may involve waiting for popular devices to become available.

When should you use real devices instead of emulators?

Use real devices when testing sensors, OEM-specific behavior, hardware performance, or other device-dependent issues.

Is emulator testing enough?

For most functional tests, yes. Use real devices for hardware features, carrier behaviour and performance numbers.

How much does a real device cloud cost?

It depends on parallel sessions. Most providers charge a subscription per concurrent session; AWS Device Farm and Firebase Test Lab also charge per device minute or hour.