Mobile QA overview

Mobile QA runs the Flutter quality checks your team already cares about — flutter analyze, flutter test, coverage, formatting, performance, accessibility — grades the result, and shows what changed since the last run. It also ships a Device Lab that drives a real emulator or simulator: install, launch, screenshot, screen-record, stream logs, and stress-test the app.

Everything runs on your machine. No accounts, no cloud, no build step.

The core idea: one grade per project

Every run produces a single, severity-weighted grade (A–F) for a project. Failing a critical check (like unit tests) costs far more than a low-severity one (like formatting), so the grade reflects real release risk, not a raw pass count. Each check rolls up into a category — Tests, Coverage, Analyze, Performance, Manual QA, and more — so you can see at a glance where the project is weak.

Runs are stored per project, so Mobile QA also tracks trends: a sparkline of recent scores and a “since last run” delta showing new failures, fixed checks, and the score change.

The main screens

Mobile QA has three top-level screens, switchable from the top nav:

Screen	What it’s for
Dashboard	The quality console — overall grade, weakest project, failing checks, and a list of every project you’ve run, sorted worst-first.
Run QA	Pick a project (or paste a path) and run the checks. Streams a live log as each runner executes.
Device Lab	Connect a booted device or simulator and check the real app: install & launch, screenshot, record, live logs, and an automated smoke.

Opening a project from the Dashboard takes you to its detail view: the grade, a category breakdown, per-test-type cards with what/why/how guidance, and the run/export/AI actions.

What you can do

Run graded QA on a Flutter project — analysis, unit/widget tests, coverage, a test-gap map, formatting, outdated deps, integration & golden tests, startup/jank/app-size performance, accessibility, and a manual sign-off checklist.
Read the grade and category breakdown, then drill into any failing check for the exact output and a concrete “how to improve”.
Auto-fix the safe stuff — dart format + dart fix --apply — in one click, then re-run.
Track trends per project with a score sparkline and a since-last-run delta.
Generate tests with AI — a prioritized test plan, a widget/unit/integration test for a file, BDD/Gherkin scenarios, an integration smoke, and a test-readiness QA summary.
Export for CI — a printable HTML report, a per-file coverage table, a grade badge (SVG), and JUnit-XML / JSON.
Drive a real device from the Device Lab — install & launch, capture screenshots and video, stream crash-highlighted logs, and run an automated smoke.

Run it

Launch the whole hub (Mobile QA starts as a child process):

node serve

Or run Mobile QA on its own:

node mobile-qa/server.js

Then open http://localhost:4113.

See Usage for a full walkthrough, or Features for everything Mobile QA can do.