Glossary
Workspace and project
A workspace groups members, projects, and shared account settings. A project contains the tests, target environments, notes, configuration, and run history for an app.
An environment is a named target URL, such as a preview app. A tunnel lets TestVibe reach a private app through a connected device.
Tests and generation
A group is a folder for related tests and can contain nested groups. A feature describes expected behavior in Gherkin or Markdown. A Gherkin scenario is one example of that behavior, written as steps.
Generated code is the saved Playwright implementation of those instructions, normally a .spec.js file. Generation writes or updates that implementation and checks it against the target app. Generation success and a later test run's success are separate outcomes.
The Assistant can help investigate, write, and propose changes. A proposal is a change prepared for review; it is not necessarily applied.
Notes are editable project knowledge organized in folders. Learned app facts can be reviewed and edited there. Keep observations specific and remove assumptions that no longer match the app.
Runs and evidence
A Playwright run executes saved test code. An Assistant run follows the written steps in a browser without requiring generated code.
A configuration defines browser and execution settings for Playwright. A device profile selects a suitable configuration or browser viewport; emulation is not a physical device.
A run contains scenario outcomes. Queued means waiting to start, running means execution is active, failed means an error was reported, and skipped means a scenario did not complete as a normal passed or failed result. It may not have started, or an Assistant scenario may have been canceled after starting; check the run details. A run can fail before it produces scenario results.
An artifact is captured evidence, such as a screenshot, video, trace, or log. Availability depends on the run configuration, completed uploads, and retention settings.
Load and telemetry
A browser load test runs browser journeys with virtual users. An HTTP load test uses k6 to make protocol-level requests. These are separate execution modes with different configuration options.
Telemetry reports measurements and errors from an instrumented app or server. It is separate from the browser evidence a functional test captures.
Storage and credits
Retention controls how long data is kept. Credits measure cloud usage; they are separate from storage capacity. Self-hosted installations use their own configured infrastructure and licensing.