Run and read results

Run a test

Every way to start a run, what happens when your plan's runs are used, and how to stop one.

From the app#

Run on a test, or Run now on its page.
Run all on a project starts every test in it. Tests that cannot start are listed with the reason.
Save and run in the composer.
Run again on a result, and Run with a new build for app tests.

A project: tests grouped by their last result, with Run on each and Run all for the project.

From elsewhere#

On a schedule.
From CI or a terminal with the command-line tool.
From an AI agent with run_test and run_project (MCP).
From the API: POST /api/v1/tests/{testId}/run and POST /api/v1/projects/{projectId}/run. Send an Idempotency-Key header so a retried request does not start a second run.

Runs at a time#

Each plan runs a set number of web runs and app runs at the same time. More runs wait their turn ("Waiting for a free slot") and start as soon as one finishes. The home page shows what is Running now.

When your runs are used#

A run starts only if your plan has runs left, counting the runs already going. When they are used, the run is not started and the app offers to upgrade or, on a paid plan, to buy 50 more runs. The test is still saved. See Plans and billing.

Stop a run#

Stop run on the run page, stop_run over MCP or POST /api/v1/runs/{runId}/stop. A stopped run is not counted.