Learn how Thunders AI works and create your first test case.
Quickstart
1. Open Thunders AI
Open Thunders AI from the left navigation to create test cases or test plans. If you are editing a test case, open Thunders AI directly in the editor.
2. Describe what to test
Write the behavior you want to cover in natural language. Focus on the expected outcome. You do not need selectors or step syntax.
3. Review and iterate
Thunders AI creates a draft test case. Review the proposed steps, then refine them in the same conversation if needed.
You can ask it to:
Change a step
Add an assertion
Update the starting page
Adjust the expected result
4. Execute the test
When the test case looks right, execute it from Thunders. If the run fails, use the run context to ask what happened or update the test.
How it works
It uses intent
Thunders AI first determines what you are trying to do. This helps separate questions from actions.
If you are | Thunders AI will |
Asking a question | Answer without changing a test. |
Creating a test case | Draft a test case from your prompt. |
Editing a test case | Apply the requested change to the current test. |
Creating a test plan | Propose a set of test cases covering a feature or workflow based on the requirements you provide. |
Reporting a bug | Use the failure context to help create a ticket. |
Investigating a test run | Analyze the run results, explain why a step failed, and distinguish between app bugs, test data issues, selector problems, or environment issues. |
It uses context
Thunders AI uses the surface you are in, such as the test case editor or a test run, to understand what you are referring to. You can still be explicit when needed, for example: "update this test case" or "use the failed step in this run."
It asks when needed
If a request is missing important information, Thunders AI may ask a follow-up question before creating or changing anything.
It keeps the conversation iterative
You can refine the result through follow-up prompts. This is useful when the first draft is close, but needs a different assertion, step order, or expected result.
Language and localization
Thunders AI follows the project's execution language for its conversational replies. If the project is set to French, Thunders AI replies in French. If set to English, it replies in English.