Changelog

The dashboard now adopts the authenticated user's browser language automatically when a matching dashboard locale is available. It currently recognizes English, Spanish, Japanese, Korean, Russian, Cantonese, Simplified Chinese, and Traditional Chinese.

You can now configure how long a test must go without a flaky occurrence before its flaky flag is automatically cleared. Previously this was fixed at 14 days. Head to your project's Automations settings to set the Cooldown period to a value that works for your team.

Tuist's pull request comments now show which specific tests failed, including the failure message and a link to the source file in GitHub. This lets you identify and fix failures directly from the PR without opening the dashboard.

Organizations can now enforce SSO authentication. When enabled, members must verify their identity with the configured SSO provider before accessing organization resources. Users can still log in normally, but will be redirected to SSO when accessing an enforced organization.

Build scans now include machine-level performance charts for CPU, memory, network, and disk usage. See exactly what your build machine was doing during the build to spot resource bottlenecks. Available for both Xcode and Gradle builds. For Xcode, run tuist setup insights to enable metric collection. Gradle builds collect metrics automatically through the plugin.

Gradle build details now show the tasks that were explicitly requested by the user (e.g., assembleRelease, connectedAndroidTest), making it easy to distinguish between different build types in CI environments with multiple builds.

Block pull requests when your app's bundle size grows beyond a configured limit. Tuist automatically creates GitHub Check Runs that compare the install or download size against a baseline branch. If the threshold is exceeded, you can accept the increase directly from the GitHub UI. Learn more in the bundle insights documentation.

See attachments associated with a given test case run directly in the dashboard to understand why your tests might have failed as quickly as possible.

You can now click on any test case run to open a dedicated detail page with everything you need to understand what happened in that run.

Tuist now supports the Model Context Protocol (MCP) over HTTP so coding agents can authenticate and interact with your Tuist data directly.

This first release includes:

• OAuth discovery and dynamic client registration for MCP clients.
• A protected MCP endpoint at /mcp with proper bearer challenges.
• Initial tools for test analysis workflows:
  • list_projects
  • list_test_cases (with optional flaky filter)
  • get_test_case
  • get_test_run
  • get_test_case_run
• An initial prompt (fix_flaky_test) that guides agents through diagnosing and fixing flaky tests.

This gives agents enough context and tooling to inspect flaky tests and drill down into run-level failure details from Tuist.