October 1, 2026
Cloud
Cloud

Context-rich test run suggestions, a custom PDF report title, and new Thunders MCP tools

TL;DR

You can now apply the context-rich suggestions from a failed test run to its test case, set your own title on a PDF test report, and give AI agents more run evidence through Thunders MCP.

FEATURE
  • Context-rich test run suggestions: A failed test run now shows a suggestion for each failed step, with the reason for the failure. You get these suggestions for every run, also for a run from CI.
  • Bulk download of test assets: Select several test assets and download them together as one zip file.
  • PDF report title: You can now type your own title for a test report PDF in the export dialog.
  • Thunders MCP: AI agents can now read run evidence, step screenshots, and the history of a test case. Agents can also get a short run status, look up many runs at once, and wait for a run to finish.
  • Bot protection warning: Thunders now warns you when the site under test runs bot protection.
  • Page error warning: A new setting makes Thunders warn you when a click causes an error message on the page.
  • Device rotation on mobile: You can now add a device rotation as a test step. The orientation setting is now in the advanced settings.
IMPROVEMENT
  • Thunders AI on mobile: The mobile settings of Thunders AI now match the run settings. You can also choose the device for mobile app browse.
  • Test assets and batch runs: The Uploaded on column of test assets now shows the owner and the date in two columns. The Batch column of Test Runs now shows the batch data file.
  • Thunders MCP: The consent page now names the app that asks for access to your account.
  • Project Onboarding: The browser Back button now moves between the steps of the new project wizard and keeps what you typed. The first app now takes the name of the project. The default environment is now called "Production".
FIX
  • Run settings: The settings that you choose in the run dialog of a single test now apply to the run.
  • Live run page: The live run page now catches up on its own when it misses updates.
  • Environments: The HTTP credentials added upon new app creation are saved to the environment.
  • Billing: A scheduled plan downgrade now keeps your add-ons after the downgrade starts.
  • Thunders MCP: A call that generates teardown steps no longer removes the main steps of the test case. The step generation tool now tells you when it writes no step, instead of failing silently.
  • App selection in Thunders AI: Thunders AI now works only with the apps that you select in the message box. It no longer asks which app to test after you select one.
  • Bug tickets in Notion: A bug ticket that Thunders AI files using Notion connector now goes to a place that your team can see, not to a private page.
  • Mobile proxy message: A mobile run with a proxy now shows a clear message: proxies are not supported yet for mobile tests.

Ready to ship faster with smarter testing?

Screenshot of the Thunders app Test Cases list, showing test sets, labels and last run status, with a label picker open