Use Application Test Execution to run Apex tests and automated flow tests from Setup and read the results in the browser. This is the manual, point-and-click path, best for ad hoc checks and namespace-scoped runs.
- Open the page
From Setup, type Application in the Quick Find box and select Application Test Execution. The page lists any recent run and a button to start a new one.
- Select the tests
Click Select Tests. Use the namespace dropdown to choose [My Namespace], a specific managed package namespace, or [All Namespaces], then check the test classes you want to run.
- Decide on coverage
Leave Skip Code Coverage unchecked for a true coverage reading before a release. Check it only when you want a faster pass or fail signal during debugging.
- Run and monitor
Click Run. Watch the live counts of passed and failed methods. Tests run in parallel by default; select running tests and click Abort if you need to cancel.
- Inspect results
Click any test to open its details, including the stack trace for a failure, or click View to see the source Apex. Results stay on the page for about 60 minutes.
Filters the test list to local classes, one managed package, or every visible namespace. The key control for package-scoped runs.
Runs the selected tests without computing code coverage. Faster, but it will not tell you whether you meet the 75 percent deploy gate.
Cancels selected tests while a run is in progress, useful when a parallel run stalls or you launched the wrong set.
Opens the source Apex behind a selected test result so you can read the code without leaving the page.
- A green pass rate is not the same as deploy readiness. Production deploys still require at least 75 percent overall Apex code coverage with all tests passing.
- Only flow tests built with automated flow testing in Flow Builder appear here. A flow with no saved test definitions will not show up.
- Live results clear after about 60 minutes. For older runs and durable inspection, use the Application Test History page.
- Parallel execution can cause intermittent failures when two tests share the same data. Fix the data setup rather than only disabling parallel runs.