Visual testing for React Native

Visual regression testing catches changes to how your app looks that nobody meant to make.

Sherlo is a visual regression testing tool for React Native. On every build, it screenshots your UI on iOS and Android simulators in the cloud, compares each screenshot against the last approved one, and shows you exactly what changed. Instead of opening the app and checking screens by hand, you review a list of differences and decide which ones are intended.

Sherlo compares each screenshot pixel by pixel against the last approved one.

Most false positives in visual testing come from screenshotting a live app - changing data, animations mid-flight, screens still loading - so Sherlo removes that at the source: every story is a fixed state with fixed data, so the same input produces the same screenshot every run.

Your app can also tell when a visual test is running, so you can freeze an animation or pause a video before the screenshot is taken. Sherlo has no model filtering out differences it thinks don't matter, so nothing gets missed.

Setup takes a few minutes - a single npx sherlo init command does the wiring for you.

It installs the SDK, connects it to your Storybook, creates the config file, and shows you the one line to add to your root component. Your stories become the test cases automatically, so there are no test files to write. See our Sherlo setup docs for the full walkthrough.

Sherlo works with Expo and bare React Native projects.

It needs a build of your app that runs on a simulator, which EAS Build produces for Expo projects.

Sherlo runs on iPhone, iPad and Pixel simulators in the cloud, from compact phones to large tablets.

All are tuned for reliable visual output, and simulators give you the most consistent environment for UI screenshots, letting you test many device configurations in parallel. You can see them all in our supported devices list.

No - with the EAS Update workflow you can reuse the same native build and test only the new JavaScript.

That turns most UI changes into a fast publish-and-test loop instead of a full rebuild. See our testing with EAS Update docs for details.

Sherlo saves hours a build, and more than it costs a month, catching bugs while they're still a one-minute fix rather than a re-release.

It automatically checks your whole UI on every device and shows only what changed. By hand, you only see the screens you thought to open, and one shared component can change dozens you didn't.

The rest is timing. Let QA find it after merge and it costs a ticket and a context switch back. Let it reach production and it costs a fix, a new build and another trip through app store review. Count the engineer hours and that one bug costs more than a month of the plan that would have caught it.

Storybook renders every UI component in isolation, which makes visual tests faster, less flaky and easier to review.

It targets how your UI looks, while E2E is built to test how your app works. A story is a UI state you defined, from one component variant to a whole screen, so Sherlo screenshots it directly across many devices, themes and languages instead of driving your app into each one.

With no flow running, screenshots don't land mid-animation or on a half-loaded screen. When a shared button changes, its own story shows the cause, so every screen it appears on is quick to confirm instead of something to investigate. E2E tests stay focused on what they're best at: validating real user paths.

Not yet - Storybook is the only integration today, but more are on the way.

Running your stories is what makes it fast: every story already describes a UI state, so Sherlo can screenshot it directly instead of navigating your app to reach that screen. If you are not using Storybook yet, it is a one-time setup that also gives you component documentation.