October brings the mabl agent into more of your workflow, from choosing which tests to run for a code change to reconciling branches. You can also connect mabl to the tools where your team already manages test cases.
New | Enhancements | Changes | Bug fixes
New
Find and run the mabl tests your code change impacts
Next time you make code changes, run the tests that matter. With test impact analysis (TIA), mabl identifies which tests your code change reaches and highlights coverage gaps. Run relevant tests before opening a pull request, review results sorted by root cause, and hand off any affected areas without coverage to test authoring. Learn more.
Coordinate agent work across your whole workspace
Now you can ask the mabl agent to take on different tasks directly from the workspace dashboard. For example, find out why last night's plan runs failed, fix the affected tests, or update every test that uses a particular URL. Learn more.
Connect TestRail, Zephyr, and Azure Boards
If your team manages test cases or work items in another tool, you can now connect it directly to mabl. Record plan run results back to your test management tool, and build new mabl tests from the cases and work items your team already wrote. Learn more about the TestRail, Zephyr, and Azure Boards integrations.
Merge conflicted branches with the mabl assistant
When a test or flow changes on master after you edited it on a branch, reconciling the two versions by hand can be tedious for longer tests. Click Merge with assistant, and mabl rebuilds each test from the changes on both master and your branch. It merges only after each reconciled test passes on your branch. Learn more.
Enhancements
Build tests from work your team already wrote
Work with the mabl agent
- Send agent-built tests back for rework, or close the task
- Clearer verification results in agent authoring sessions
- See when an agent made a change
Run plans your way
- Choose environments when you run a plan
- Use different credentials for a single plan run
- Run specific plans by ID from your pipeline
Do more from your AI client
- Review performance test results
- Manage environments and look up workspace members
- Find plans by label
- Find tests by test case ID
- Follow your mabl work live in Claude Code
Authoring and triage
- Set an HTTP header from a test step
- Wait for an element's value before asserting
- Jump from the Results summary to the failed step
- Archive, restore, and delete custom failure reasons
Changes
Capturing step trace is now opt-in
Cloud runs no longer keep step traces by default. Performance metrics such as speed index aren't affected. To keep capturing them, turn on Retain step traces in a plan's additional browser settings. Learn more.
Support for iOS 27.0 and discontinuing iOS 17.2
Mobile cloud training and execution now supports iOS 27.0 with iPhone 17 and iPhone 18 Pro. iOS 17.2 is no longer supported. Learn more.
Post-quantum cryptography for connections to mabl
Connections to the mabl web app, Desktop App, CLI, and REST API now use post-quantum cryptography, which protects your traffic against "harvest now, decrypt later" attacks. Learn more.
Bug fixes
General
- Fixed an issue where removing a user from a workspace failed in both the web app and the mabl API.
- Fixed an issue where the workspace credit limit warning appeared even when the workspace limits added up exactly to the account total.
- Fixed an issue where DataTable rows went missing when parallel test runs created scenarios at the same time.
- Fixed an issue where the Xray test case ID filter on the Tests page showed no IDs when the page first loaded.
Test authoring
- Fixed "test or flow is too large" errors when saving a test whose flows use large values from earlier steps, such as HTTP responses.
- Fixed an issue where the Save button for an element assertion stayed disabled when the expected value was blank.
- Fixed an issue where the Expected value for a CSS assertion created from Find elements no longer filled in automatically with the element's current value.
- Fixed an issue where the Trainer stopped responding when the JavaScript snippet editor was closed with the operating system's window controls.
- Fixed an issue where the AI prompt disappeared from the JavaScript snippet editor.
- Fixed an issue where Find by XPath couldn't be removed from an assertion step in a mobile test once it was configured.
- Fixed issues in the iOS mobile Trainer, including crashes on steps that select an element, a cloud device that stopped responding after a session was extended, and a missing crosshair for visual-area Create variable steps.
Test execution
- Fixed issues with Safari cloud runs, including crashes and slow runs on certain pages, failing wait steps, and "Connection terminated unexpectedly" errors on the first navigation to some sites.
- Fixed an issue where requests that followed a redirect to another domain were blocked by CORS in cloud runs.
- Fixed an issue where pages intermittently failed to load in cloud runs with "This site can't be reached" after a test had loaded large files from the same site.
- Fixed issues where PDFs embedded in your application, such as in a modal or an iframe, didn't display on the page during test runs.
- Fixed issues with email steps, including open email steps that failed in the Trainer and local runs when an HTTP basic auth credential was applied, and variables in the To field that weren't replaced with their values.
- Fixed an issue where custom finds that used
nth=didn't select the same element consistently. - Fixed an issue where an API step that used
pm.sendRequestin a post-request script was reported as skipped and its assertions were ignored. - Fixed an issue where JSON body assertions in API tests produced a deprecation warning.
- Fixed an issue where tests in a plan stayed queued and then stopped with "the test runner stopped responding."
- Fixed an issue where staged plan runs never moved past the first stage after an API test finished.
- Fixed an issue where a performance test ran against a different environment from the one configured.
Test output
- Fixed two issues with test result PDF exports: exports from the mabl Desktop App failed with "File download error," and plan names with Japanese characters were cut off.
- Fixed an issue where selecting a Test metrics checkbox on performance test results crashed the page.
- Fixed an issue where performance runs that mabl stopped showed only Stopped, with no reason given in the app or the mabl API.
mabl API
- Fixed issues with API tests created through the mabl MCP, which could fail to save with a "Bad Request" error or leave variables unresolved.
Integrations
- Fixed issues with reporting results from the Xray sync integration, including loops, screenshots, and the Actual Result field.
- Fixed an issue where setting up Azure Boards from the mabl Desktop App left users on the Microsoft sign-in page with no way back to mabl.
Subscribe to our monthly newsletter
Interested in getting monthly updates sent directly to your inbox? Click here to subscribe to the monthly product newsletter.