For teams whose suite already runs in Certyn once, and who want it every night (or at any fixed time) without a CI job.
Your setup
- Tests in a repo, with repo access saved (see Run your tests in the cloud), or a Certyn suite under Tests.
- A test environment, for example staging.
- You want a run at a fixed time, in your time zone.
Say this to Certyn
Run our Playwright suite against staging every night at 2am Berlin time.
What Certyn does
- Checks that it has repo access and the saved repo and test command. If the suite has not run in Certyn yet, run it once from the chat first.
- For your own tests, creates an agent with the same goal: clone the repo, install, run the suite against the environment, report. It shows under Agents. It asks you to confirm.
- Sets the schedule for staging: cron
0 2 * * *, time zoneEurope/Berlin. You approve the settings change. You can also change the schedule later on the agent's or suite's settings. - Each night the agent starts a run by itself.
A Certyn suite (test cases with a tag), or a TestRail, Zephyr, Xray or Azure Test Plans suite, gets a schedule the same way: run the Checkout suite on staging every weekday at 6am New York time.
The goal per framework
Certyn writes the goal from what worked in the first run. A typical one:
| Framework | The run command in the goal |
|---|---|
| Playwright | npx playwright test |
| Cypress | npx cypress run |
| Selenium (Java) | mvn test |
| Selenium (Python) | pytest --junitxml=junit.xml |
| Selenium (C#) | dotnet test |
| WebdriverIO | npx wdio run wdio.conf.ts |
| Postman (Newman) | npx newman run postman/api.json -e postman/staging.json |
| Hurl | hurl --test api-checks/*.hurl |
| k6 | k6 run loadtests/checkout.js |
What you get
- A run report each morning, with failures grouped by cause.
- New findings only for new product bugs.
- The history of every night's run on the agent.
Limits
- A scheduled run has 60 minutes. For a longer suite, split it into two agents, for example by Playwright project or folder, each with its own goal.
- To check something at a fixed local time (for example just after midnight in Sydney), the schedule's time zone is the way: a run can change the browser's time zone, not the clock.
- GitHub App rules (pull request, merge, push) and the CI Runs API start only Certyn suites of test cases, not an agent's goal. For a run after each deploy, turn on after each deploy for the same agent. See Run a suite after each deploy.
- The Regression suite has no fixed schedule: the weekly Agenda decides when it runs.
- Each run uses plan time for as long as it takes.