Skip to content
Marriska Docs

Quickstart

Let’s get a test running. Three minutes, no setup, nothing to install.

Before you start

Section titled “Before you start”

You need:

  • A Marriska account (sign up — starts on a 14-day Pro trial, no card; drops to the Free plan afterwards)
  • A modern browser

That’s it. Tests run in the cloud on Free, so there’s nothing local to install for this tutorial.

Steps

Section titled “Steps”

  1. Sign in and open Tests

    After signup you’ll land on the Dashboard. Click Tests in the top nav (or the bottom bar on mobile).

  2. Create a new test

    Click New Test in the explorer sidebar (or right-click the explorer and pick New Test from the context menu). Give it a name and confirm.

  3. Describe the test in plain English

    The editor opens with a description box. Paste this in:

    Open https://example.com
    Check that "Example Domain" is visible on the page
    Take a screenshot

    Click Generate Steps. The parser turns each line into a structured action — you’ll see the steps appear underneath the description.

  4. Run it

    Click Run Test. Pick Chromium (or all three browsers if you want). Step events stream back live — green checkmarks, durations, and screenshots fill in as Playwright works through the steps.

  5. Open the report

    When the run finishes, the Reports page has the row you just created. Click it for the full per-step view: status, screenshots, timing, and a stable URL you can share with anyone in your org.

What just happened

Section titled “What just happened”

You wrote three sentences and Marriska:

  1. Translated each into a Playwright action (page.goto, expect().toBeVisible, page.screenshot)
  2. Spun up a real Chromium browser in the cloud
  3. Ran the steps, captured screenshots at each one
  4. Persisted everything as a shareable report

No Selenium grid, no WebDriver setup, no Node install.

Waiting for slow operations

Section titled “Waiting for slow operations”

Some flows genuinely take a while — report generation, imports, long backend jobs. Two ways to say that in plain English:

  • Preferred — wait for a signal: “Click Generate and wait until ‘Report ready’ appears, up to 3 minutes.” The test continues the moment the text shows up, and fails with a clear reason if it never does.
  • Fixed pause: “Click Apply and wait 3 minutes.” Sleeps exactly that long (converted to seconds automatically; capped at 10 minutes). Use it only when there’s no visible signal to wait for — a fixed sleep is slower when the job finishes early and still fails if the job runs late.

Both are also available as manual steps in the editor: Wait until text appears (with a max-seconds budget) and Wait (seconds).

Next steps

Section titled “Next steps”