QA Flows Overview
QA Flows let you generate Journeys from natural-language requirements and run them against your application in a real browser. Instead of writing tests by hand, describe what you want to test and let Marketrix build executable Journeys for you.
How QA Flows work
Section titled “How QA Flows work”- Create a QA Flow — Provide a description or upload a PRD. Marketrix turns it into structured Journeys. Each test flow is called a QA Flow and contains a set of Journeys plus the personas that perform them. A flow is saved in one step — there’s no draft or review-before-save state.
- Edit the Journeys — Open the flow’s detail page to refine its Journeys and personas at any time.
- Run the flow — Execute the Journeys against your application in a real browser. Each execution is called a QA Run.
- Review results — See which Journeys passed, which failed, and which were indecisive, and why. Every Journey is evaluated automatically and given a verdict.
Key features
Section titled “Key features”- AI-powered generation — Describe what to test in plain language, or upload a PRD; Marketrix writes the Journeys.
- Real-browser execution — Journeys run in Chrome across selectable viewports (desktop, tablet, mobile) — no scripting required.
- Automatic evaluation — Every Journey is judged against its expected outcome and given a passed / failed / indecisive verdict.
- Run history & reports — Track every QA Run with passed/failed/indecisive counts and a per-Journey report that includes screenshots and the steps Marketrix actually followed.
- Multiple viewports — Run the same flow across desktop, tablet, and mobile viewports and compare results side by side.
- Editable flows — Open a QA Flow’s detail page to edit its Journeys and personas at any time.
Verdict categories
Section titled “Verdict categories”Each Journey in a QA Run receives one of these outcomes:
| Verdict | Meaning |
|---|---|
| Passed | The application reached the expected outcome and the followed steps aligned with the planned steps. |
| Failed | The expected outcome was not reached. This usually points to an actual bug in the application. |
| Indecisive | The agent could not conclusively pass or fail the Journey; you resolve it, and the resolution is remembered per application. |
Every verdict carries a short, human-readable reason explaining why the Journey passed, failed, or was judged indecisive.
Next steps
Section titled “Next steps”Creating QA FlowsGenerate Journeys from a description or a PRD.
QA RunsRun flows, read verdicts, and download reports.
SimulationsThe browser automation QA Runs are built on.