Skip to content

Add Them to the Repository

2 min read

Adding is the only step that writes anything. Up to that point a run lives entirely in the modal — close it and nothing has changed.


The footer carries a destination picker, which opens on the folder you had selected when you started.

  1. Pick an existing folder from the list, or

  2. Select the new folder icon and type a name — it is created inside the folder shown beside it.

  3. Select Add test cases.

If you leave it on Repository root, that is where they go.

Folders are matched by name when you create one, so typing the name of a folder that already exists puts the cases in that folder rather than creating a second one with the same name.


An accepted draft becomes an ordinary test case. Specifically it gets:

  • its own test case code, from the project’s sequence;
  • the folder you chose;
  • links to the requirements it covers, so it counts toward coverage immediately;
  • your project’s custom field defaults, the same ones the manual create form applies;
  • an entry in the audit log and its own change history.

The status is active. There is no draft, review or approval state — ticking the drafts in the modal is the review, and nothing is waiting on anyone afterwards.

If your project has a required custom field with no default, the cases are still added and a notice names the field, so you know what to fill in. A required field never blocks a batch you have already reviewed.


Three things mark a generated case, so “which of these did the AI write?” is always answerable.

A small AI badge sits beside the title, everywhere test cases are listed and on the test case itself. Hand-written cases show nothing at all.

On a test case’s detail view, a read-only panel shows where it came from:

  • the scenario that asked for it,
  • the lens that scenario came from.

The scenario used to be part of the generated title. It is a property now, so titles read as titles.

In the Repository, filter by Source to see only AI generated or only Written by hand cases — useful for spot-checking a batch, or for seeing how much of a suite is machine-drafted.


Generated cases behave like any other test case from this point. They can be edited, moved, added to suites, run in executions, and counted in coverage and reports. The AI tag is a fact about where the case came from, not a state — it stays, and nothing depends on it.