Desktop app
An installed QA app for managing projects and execution evidence on your computer
reopt QA is an installed desktop application. The directly distributed installer includes an execution engine and browser runtime, so you can run tests and inspect their results from the app.
Installation and prerequisites
Local automation currently uses the directly distributed macOS app. Check supported systems and installer availability on the download page. The Mac App Store distribution does not provide local automation.
You still need a target URL, test accounts, and test definitions. AI features require separate model provider authentication. Follow Getting Started for your first run.
Work in the app
| Task | Features |
|---|---|
| Understand product structure | Collect pages and inspect connections and page results in the service map |
| Define coverage | Manage tests, the feature catalog, environments, and personas |
| Check actual behavior | Run tests with the intended environment, persona, and browser |
| Investigate failures | Review reports and available screenshots, logs, and traces |
| Verify improvements | Rerun tests, compare results, and inspect analytics trends |
| Repeat checks | Process interval schedules while the local engine is running |
The main navigation is Overview, Service map, Verification, Improvements, and Changes (개요 · 서비스맵 · 검증 · 개선 · 변경 내역). Accessibility, performance, visual regression, and debugging evidence are available when the corresponding checks and collection are configured.
AI and detail panels
AI chat opens a shared panel with context from the selected page, feature, flow, or improvement. Review or exclude attached context before sending. Open task history and configuration through Settings → AI tasks, chats, and settings (AI 작업·대화·설정).
At smaller widths, opening AI hides some detail panes; closing it restores the selected detail. Switching project tabs restores each project’s screen selection within the current app session.
Separate projects and conditions
Use a project for each service and manage its environment URLs and tests. Keep staging and production distinct. Compare runs with matching environments, accounts, and browsers. Registering multiple projects does not automatically test all of them.
External AI tools can connect through the MCP bridge. The bridge currently comes from the repository CLI and is not included in the app installer.
Stored data
The directly distributed app stores user data in agent-data under its local application data directory. The separate runtime directory contains replaceable execution components. Fully quit the app and let running work finish before copying its data directory.
Local storage does not mean that all features avoid external transmission. AI tasks can send required data to the configured model provider. See data flow and costs and privacy.