v2026.11.12
TL;DR
Watch Thunders AI browse your app live in chat, save execution settings for test sets, and repeat runs with their previous settings. Other updates include mobile swipe steps, native pickers, and more detailed RGAA reports.
FEATURE
- Watch Thunders AI browse live: Open Watch live from a browse action in chat to follow Thunders AI’s browser. The preview is view-only.
- Execution settings for test sets: Save execution settings in a test set’s Execution settings tab and reuse them for future runs.
- Repeat a run with its settings: Re-execute using the settings of the run being repeated, or adjust them before starting.
- Bulk schedule actions: In a test set’s Schedules tab, select multiple schedules to run, pause, resume, or delete them together. Restrictions still apply to resuming system-paused schedules.
- Mobile tests on virtual devices: Run mobile tests on Android emulators and iOS simulators using compatible builds.
- Mobile swipes and native pickers: Add swipe steps targeting an element or the screen. Thunders also supports selecting options and dates in native Android and iOS pickers.
- Custom browser resolution: Set custom browser dimensions, up to 15 million pixels in total.
- Create multiple test cases: Ask Thunders AI for up to five test cases in one request. It browses and creates them one at a time, using a fresh browser context for each case.
- Find Xray Cloud tests by topic: Ask Thunders AI to find relevant Xray Cloud tests by searching their titles, descriptions, and steps.
IMPROVEMENT
- Edit test cases through MCP: Your AI client can update a test case’s status, name, description, environment, and persona.
- Test-entities links in AI responses: Open test sets, test plans, test cases directly from links in Thunders AI responses.
- More detailed RGAA reports: Web RGAA PDF reports include suggested fixes where available and images of affected elements when capture succeeds. When an element cannot be captured, the report shows the reason instead.
- Faster failed assertions: Thunders does not retry an assertion when the value does not match. When the page did not change, Thunders gives the same result again without a new AI request.
- Data rows in run reports: See which data-file row each run used in run details, run lists, and PDF reports.
- Visual-only suggestions for canvas steps: Thunders recommends Visual only mode for steps targeting a canvas. You choose whether to apply it.
- Loop re-runs: A re-run of a loop continues in the open browser of the failed run. The run card shows where each loop attempt starts.
- Certificate warnings: In Chromium, pages with invalid certificates show a “Not secure” warning during recording and when reviewing affected run steps, with details about the certificate error.
- Expected API errors through MCP: Configure API steps through MCP to expect status codes such as 4xx or 5xx, so tests can verify rejected requests without treating the expected response as a failure.
- Device operating system: Mobile run pages show the device’s operating system and version when available.
- Step keyboard shortcuts: Copy selected test steps with Cmd/Ctrl+C, or open the delete confirmation with Backspace or Delete.
- Accurate page-stability settings: Project loading settings display the actual defaults and post-action timing values. This corrects the displayed settings, not runner behavior.
- Clearer project setup: After sign-up, setup starts at the app step with a default project. The empty My projects page provides a clearer prompt to create a project.
FIX
- Slider verification: Slider actions that leave the value unchanged no longer incorrectly pass.
- Accented characters and conditional steps: Fixed key presses involving accented characters. Conditional steps no longer retry
- Mobile tap targeting: Android tap targeting avoids elements hidden behind overlays, and the keyboard is dismissed when it blocks the target.
- iOS preview rotation: The iOS live preview displays upright after a device rotation into landscape.
- Schedule settings: Custom headers and Appium capabilities are saved correctly.
- Run view: The execution panel follows the step that runs. The last step stays active until the run ends. The step list stops its auto-scroll while you use it.
- Xray Cloud test runs: Thunders AI can read Xray Cloud test runs without the previous defects-query error.
- Defect destinations: Thunders AI no longer offers Xray as a destination for filing defects. Connected issue trackers such as Jira remain available.
- Browse-setting preservation: The browse switch is preserved when a conversation moves into a test case.
- Targeted AI edits: Targeted edits no longer regenerate the whole test case or flatten untouched Discovery sub-steps.
- Dark-mode controls: Improved the contrast and interaction states of device-setting controls and number-field arrows.
- File upload and country search: Fixed the file-upload interface. Country-picker searches show the first match rather than retaining the previous scroll position.
- Live connection recovery: Fixed recovery when the initial live connection fails.
- Visual Compare: Cropped images in the difference list fit the available space without distortion.
Migration steps
- Pull the new images and redeploy, starting with the API Service (it applies any database updates automatically).
upgrade urgency
Critical
Ready to ship faster with smarter testing?
