Oct 19, 2026
2 min
Changes released on Oct 19, 2026.
New
- Improved design validation for web and mobile. Design validation now has a refreshed interface, a selector for web and mobile targets, and better element matching, scaling and capture. Scenario flows can be simplified with AI assistance.
- Azure DevOps and Jira tickets from Playwright runs. Create Azure DevOps work items and Jira issues from Playwright results, with attachments. Linked tickets appear in a dedicated tab of the report, and ticket creation now runs in the background.
- Choose the authoring model and proxy per scenario. In the Playwright wizard each scenario can pick its own authoring model and decide whether to use a proxy. The results screen shows which model and providers were used, and rejected models now come with a clear explanation.
- Conditional blocks, dismiss steps and API conditionals. Playwright scenarios can contain conditional blocks that are shown as skipped when their condition never appeared, plus a dismiss step. API tests gain conditional wrapping, retry and extraction controls.
- Auto-heal in the run wizard. Runs can enable auto-heal from the wizard. Heal badges, a heal panel and a live view appear on the run and report pages, and the heal panel now shows the value heal class.
- Unique values in test data. Test data rows can mark a value as unique per run with a placeholder, toggled from the row's Actions column. Saving a data set checks the unique values and explains any rejection.
Improvements
- Scenario status column and server-side paging. The Playwright scenario list shows one clear status per scenario (Passed, Failed and others) and loads one page at a time with search, status and folder filters. Page, filters and folder are kept in the URL and restored when you return from a scenario.
- Faster lists, reports and dashboards. Reports, run histories, schedules, dashboards and accessibility results load in pages and are noticeably lighter on large accounts.
- Clearer live runs and blocked-run explanations. The run page now follows the engine live, with steps advancing as the test runs and the engine's time per step on the results card. A blocked step shows one verdict with the site's answer and collapsible engine detail, and steps that were never reached are listed as not run.
- Recorder improvements. Recorder AI assistance is now chosen per scenario in the wizard instead of a company-wide switch. The live recording list shows readable step labels, and the Recorder v2 address bar follows the page being recorded.
Fixes
- Stronger access control and data isolation. Closed cases where report lists, scenario exports and version endpoints could expose another tenant's data, tightened authorization on approval endpoints, and hardened sharing of reports, data sources and dashboards. Two-factor secrets are no longer returned in API responses.
- More reliable ticket creation. Tickets are created once per execution, list items in descriptions keep their nesting in Jira and HTML, and a renamed Azure DevOps project keeps receiving tickets. Cancelled runs stay recognisable in tickets.
- Corrected dashboard metrics and quotas. Dashboard metrics were corrected and widget and retained-report quota limits are now enforced accurately.
- List and run page fixes. A scenario list refresh that returned after you changed page no longer shows the old page, and collapsed folder groups no longer hide rows on other pages. The run row menu now closes correctly, a finished step shows its tick, and deleting a run removes its whole retry chain.