Your first test, end to end

The fastest way to understand Testsnag is to build one real test against a site you know. Pick something short and important: a homepage that must load, a search that must return results, a sign-up form that must accept an address. Avoid starting with your most complicated flow, because the goal here is to see the whole loop rather than to cover everything.

12345

1. 1Add a step by hand when you know exactly what you want.
2. 2Describe the step in plain English and let the AI find the element.
3. 3Drive the live browser yourself and have your clicks recorded.
4. 4Finish with an assertion. Without one, the test cannot really fail.
5. 5Run it once from here to see the whole loop end to end.

The loop

1. 1

   Create the test

   Open Tests in the sidebar and click New Test. Name it after the journey, pick the Web type, and give it the URL the journey starts from.

2. 2

   Record the journey

   The recorder opens a live browser. Click Start Recording and use the site exactly as a customer would. Every click and keystroke becomes a step.

3. 3

   Assert the outcome

   At the end of the journey, add an assertion for the thing that proves it worked: the confirmation text, the new URL, the element that should now be visible.

4. 4

   Replay it once

   Replay the steps from the recorder before you leave. This catches the flakiest problems immediately, while the context is still in your head.

5. 5

   Run it for real

   Click Run Test. Testsnag opens a fresh browser, walks the steps, and reports a pass or fail with screenshots for each step.

6. 6

   Put it on a schedule

   Open the test settings, switch to Scheduled Runs, and pick an interval. A test nobody remembers to run is not really monitoring anything.

If the first run fails

That is normal and usually informative. The most common causes are a step that ran before the page was ready, a selector that matched something during recording but not on a fresh load, or an assertion written against text that changes between visits. Open the run, find the first failed step, and look at its screenshot: it shows you exactly what the browser saw at the moment it gave up.

The troubleshooting articles cover the recurring causes in more detail.

Then what

Once one test is green and scheduled, add two or three more covering your other critical journeys, then group them into a suite so they run together and notify you as a unit. Resist the urge to cover everything at once: five reliable tests you trust are worth far more than forty that cry wolf.