v2026.09.24
TL;DR
Thunders AI now knows which environment and app it works in, your tests can start already signed in with Authenticated Profiles, and you can now see every failure that self-healing prevented.
FEATURE
- Environment-aware Thunders AI: Pick an environment and an app above the Thunders AI message box. Every message after that resolves variables, credentials, and addresses from your choice. Web browse opens the right environment with its real execution settings, and each executed step records which environment and settings it used.
- Failures Avoided with self-healing: When a UI change breaks a selector, Thunders repairs it during the run. The test passes instead of failing for a reason that is not a bug in your app. You now see each of these saves: the healed step in the run, the count in the run strip, and your Failures Avoided score and its trend in Analytics.
- Authenticated Profiles: Capture a logged-in browser session once and reuse it across your tests. A Capture profile step saves the browser state after the login. An Inject profile step restores that state in a fresh browser, so the test starts already signed in. One step now replaces the login steps or the shared login module in each test. Thunders checks each injected session and reports an expired or rejected session at the inject step.
- Bulk-assign test cases to test sets: Select several test cases in the list, then add or remove them from one or more test sets in one action.
- Proxy support for test runs: A run's advanced settings now include a proxy field. Paste your provider's connection string and the run routes its traffic through it.
- Turn a ticket into tests: Thunders AI now offers a "turn a ticket into tests" prompt. Pick a connected ticket source, or paste the ticket content directly, and Thunders AI drafts the test case.
- PDF report overview page: A test report PDF now opens with a summary page that states the counts for the run it covers.
IMPROVEMENT
- Copy a member's email: The organization members list now lets you copy a member's email address in one click.
- Duplicate an environment: The environment list now offers a duplicate action.
- RGAA conformance rate: The accessibility audit report now qualifies its conformance rate, and fixes the wording of an item it cannot automate.
- Current Gemini model guidance: The startup check now warns you when you configure a Gemini model that Google has shut down or is retiring. The documentation and examples now use gemini-3.1-pro-preview and gemini-3.1-flash-lite.
FIX
- Visual Compare accuracy: Visual Compare now draws its difference boxes on the correct rows, even after the page shifts vertically between the two captures.
- Cleaner test authoring: Choosing an option from an autocomplete field no longer adds an extra step. Dismissing a popup while recording now creates a conditional step, instead of one that always runs. Clicking empty space no longer adds a step, and recording now captures interactions inside embedded frames.
- Steadier test runs: A dropped connection to the browser driver now reconnects on its own instead of failing the run. A download that never finishes now fails clearly instead of hanging.
- Test set duration: A test set's duration now reflects the true combined time of its runs, instead of adding up time that two runs already shared.
- Test sets with the same name: When two test sets share a name, picking one of them now selects only that one, not every test set with that name.
Ready to ship faster with smarter testing?
