Release notes··4 updates

    Diff-anchored PR verdicts, cross-bot coverage, and faster dashboard

    4 updates shipped on July 1, 2026.

    1. Highlight

      Featured this release

      PR review verdicts now require evidence the diff caused the failure

      A failing check no longer blocks your PR unless the review agent can trace the failure back to something in your diff.

      A failure only blocks a PR review when it's confirmed to have been introduced by the diff. Pre-existing bugs, unrelated failures, and flakes move into a folded "pre-existing issues" section instead of sitting in the blocking list, so the blocking verdict reflects what your change actually broke.

      # Direct link

    2. Update 02 of 04

      Faster page loads across the dashboard

      Dashboard, Test Cases, Results, Issues, Insights, and Coverage pages now show instantly with a lightweight skeleton while content streams in behind it.

      Navigating between pages no longer waits on the slowest piece of data – headers and layout appear immediately, and tables and charts fill in as they're ready. Run history and revision panels also load lighter, since they now load only when you open them.

      # Direct link

    3. Update 03 of 04

      PR review testing covers mobile and tablet without manual setup

      When a PR touches mobile, tablet, or responsive behavior, the review agent now creates a matching device preset instead of skipping the test.

      If your project doesn't yet have a viewport preset for the device a PR review calls for, QA.tech creates one automatically (reusing any existing preset first), so cross-device coverage doesn't depend on someone setting it up ahead of time.

      # Direct link

    4. Update 04 of 04

      PR review bot cross-checks other review bots' findings

      If Claude or Cursor Bugbot already flagged a UI-testable issue on your PR, QA.tech's suggested tests now cover it too.

      When drafting test suggestions for a PR, the review bot now reads through existing comments from tools like Claude and Cursor Bugbot, looking for UI-testable bugs, edge cases, or regressions they've already flagged. Those findings get folded into the tests it proposes, so your suggested coverage keeps pace with what your other review bots have already caught.

      # Direct link

    Your team moves fast. Can your testing keep up?

    QA.tech agents test your product autonomously, so moving fast never means shipping broken. See how it works in a 30-minute demo.

    Get a demo