Zerocheck shows the failing step, screenshots, console errors and every attempt. A pass after a supported transient retry stays labeled flaky.
Failing steps with screenshots
Retry results stay labeled flaky
Assertion failures are not retried
Rerunning a failure without investigating teaches the team to ignore CI.
A locator change, an assertion failure and a service error need different responses.
The original failure should remain visible even if a later attempt passes.
A developer sees a red job, reruns it and loses the context needed to understand why it failed.
Open the run, inspect the failing action and expected outcome, and compare the step screenshots. Recovery and flaky labels remain part of the result.
Read the failing step and the exact test revision.
Check the screenshot, console output and failure type.
Fix the product, test data or test description based on the evidence.
Use --fail-on-flaky in CI when a flaky blocking test should fail the workflow.
Paste your QA checklist. Zerocheck drafts editable browser tests, runs them in your app, and shows the result of each item before you save it.
Zerocheck keeps the exact test revision, step screenshots, errors and attempts together so an engineer can explain a browser test result.
Start with a browser test in the web app, then use the optional CLI to run repository YAML in GitHub Actions and post a result on the pull request.
No. A transient AI-service error can be retried once in the same browser. Assertion failures are not retried.
It provides failure details and browser evidence. Engineers still decide the cause and the fix.
Zerocheck shows the failing step, screenshots, console errors and every attempt. A pass after a supported transient retry stays labeled flaky.
Get a demo