QaAgent is live: local-first browser QA automation

[BAKUGOS1]

QaAgent is live as a local-first TypeScript + Playwright QA agent for real website testing.

What it can do today:

• Run browser QA from a URL or task file
• Capture screenshots, traces, browser state, console errors, and network errors
• Generate Excel-first QA reports with coverage truth
• Run in Codex/no-API mode or Groq/API mode
• Keep destructive actions blocked by default

The goal is simple: make website QA evidence-rich, safe, and practical for developers/product teams.
Useful links:

• README: https://github.com/BAKUGOS1/QaAgent#readme
• Sample report: https://github.com/BAKUGOS1/QaAgent/blob/main/docs/SAMPLE_REPORT.md
• Launch kit: https://github.com/BAKUGOS1/QaAgent/blob/main/docs/LAUNCH_KIT.md
• Plugin guide: https://github.com/BAKUGOS1/QaAgent/blob/main/docs/PLUGIN_INSTALL.md
• Good first issue Add more ready-to-run task templates #1: Add more ready-to-run task templates #1
• Visual regression issue Add opt-in visual regression checks #2: Add opt-in visual regression checks #2
• Report gallery issue Create a sample report gallery #3: Create a sample report gallery #3
• Selector healing issue Improve selector healing for common SaaS UI patterns #4: Improve selector healing for common SaaS UI patterns #4

[BAKUGOS1]

Maintainer note: I built QaAgent because I wanted website QA runs to produce real evidence, not just a loose checklist. The first goal is simple: give developers a local browser agent that can test flows, capture proof, and produce reports that are actually useful in product work.

If you try it on a real public site or staging app, share what worked and what felt missing. That feedback will decide the next few improvements.