In 2025, we remained committed to delivering practical AI solutions and strengthening the essential platform capabilities - maintenance, execution, infrastructure, integrations - that you rely on every day. 52 weeks and 63 releases later, here’s a look back at all the noteworthy enhancements from the past year.
The mabl Test Creation Agent
In June, we introduced the Test Creation Agent. Using your provided natural language prompt, the agent builds out browser, API, and mobile tests based on your intent and uses semantic search for flows and tests to ensure consistency with your existing testing conventions.
mabl MCP for AI workflows
When you set up the mabl MCP in your preferred AI client, you can access mabl’s testing tools right where you work. Introduced in July, we have been steadily expanding the mabl MCP toolkit to support different agentic workflows, including the ability to create and run tests, review results, and integrate with other MCP tools.
Visual finds and element descriptions
To ensure more reliable finds, you can now use AI to find elements based on visual characteristics:
- Visual taps and clicks - instead of targeting elements based on text-based attributes, visual find uses GenAI to tap or click on elements based on their visual characteristics. This approach works especially well for difficult-to-target elements like SVGs, canvases, and images.
- Element descriptions - for all recorded mobile tap steps, mabl captures an element description. During execution, if native find methods aren’t enough to reliably target the element, mabl uses a generated visual element description to target the element instead.
- Language settings for generated descriptions - if your working language isn’t English, you can change the default language for generated visual find and element descriptions.
GenAI Assertions for file downloads
In October, we introduced GenAI Assertions for file downloads to help you cover more end-to-end testing scenarios. With GenAI Assertions, you can validate the content of a downloaded file as well as compare it to a screenshot of the page it was downloaded from. Supported file types include CSVs, images, PDFs, audio, and video.
Auto Test Failure Analysis (Auto TFA)
Several enhancements were also made to Auto TFA, so you can quickly understand failures and reduce overall time spent on resolving issues.
- Auto TFA for plans - in addition to tests, mabl also now generates failure analyses for failed plan runs
- Jira issue pre-fill - Auto TFA automatically generates a summary and description for Jira tickets created in mabl
- Access via mabl API - to support different reporting workflows, the test run summary and plan run summary endpoints now include a “failure_analysis” object in the response with details on why the plan or test failed.
Xray results reporting
Teams that use Xray for test case management can now use the Xray sync integration to publish mabl test run artifacts to an associated Xray test case.
Unified reporting
With Unified Reporting, you can consolidate your browser test results from cloud, local, and Playwright runs into a single location in the mabl app. Introduced in April, Unified Reporting can help your team identify trends, enhance collaboration, and present a more holistic view of the quality of your app.
API testing enhancements
To better support API testing at scale, we made the following enhancements to the API Test Editor:
- Reusable flows for API tests - encapsulate commonly used sequences for reuse in other API tests
- OAuth 2.0 support - automatically generate and send access OAuth 2.0 tokens to your API endpoints
- Import curl requests into API tests - quickly convert a request from the command line to the mabl API Test Editor
- Disable steps - skip requests with known issues or to isolate an issue
mabl Link enhancements
Expanding on our commitment to secure, enterprise-grade infrastructure, we delivered several enhancements to mabl Link:
- Run performance tests over mabl Link - if technical constraints prevented your team from using static IP addresses, you now have the option to run performance tests over a mabl Link-enabled environment.
- Connection filters - if your organization requires more granular network control, you can now add a connection filter to a Link Agent configuration file to restrict outbound traffic
- Auto-reload - update settings in the Link Agent configuration file without having to restart the agent
Plan execution overrides
To support more a wider range of testing scenarios, we added the ability to override the following default settings for individual tests within a plan execution:
- Credentials - configure credentials either as a default for the entire plan run or as a test-specific override
- DataTable scenarios - assign specific DataTable scenarios to a specific test during a plan execution
Faster run times on new cloud runs
In July we introduced new browser runners for faster test startup and quicker feedback cycles.