Start free Get a demo

With coding agents opening pull requests faster than ever, the slow part is no longer writing the change. It’s proving the change works and didn’t break anything else. Running UI tests locally means waiting on the few simulators your machine can handle, so a lot of developers fall back to clicking through the app by hand and miss scenarios.

Run the whole suite in parallel

Revyl runs your regression workflow on cloud simulators in parallel and posts the result as a check on the pull request, with screenshots of the change. In the video, four tests run at once, and the failing check links straight to the run.

Four tests ran side by side, and you can play back each one.

Debug any step

Open a single test to see a summary, its key moments, and every instruction and validation alongside the video of that step. There are no selectors to maintain, because each step is written in plain language and checked by a vision model. The performance and state tabs show what the app and the network were doing during the run, which helps when you’re chasing a slow screen.

Every step of the add-to-cart test, with the screen the validation checked.

Try it on your app

  1. Write a few tests for your critical flows and group them into a workflow.
  2. Run the workflow on your pull requests so the result shows up as a check.
  3. Open any failing test to watch the step that broke.

Questions

How many Revyl tests can run at once?

Revyl runs tests on cloud simulators in parallel, so a workflow of five, ten, or twenty tests finishes in about the time of its slowest test instead of one after another on your machine.

Do Revyl tests use selectors?

No. Each step is written in plain language and checked with a vision model, which avoids the flakiness of selector-based tests.

More guides

View all