I built an open-source AI QA engineer that tests your web app in a real browser

I've been building QA Panda — an open-source AI QA agent that lives inside VSCode.

You say "test the login page" and it launches headless Chrome, navigates your app, clicks through flows, finds bugs, takes screenshots, and gives you a professional QA report.

How it works

1. Open any project in VSCode
2. Open QA Panda and type what you want to test
3. The AI agent launches a real browser, creates a test plan, and executes it step by step
4. Bugs are logged with screenshots and severity levels
5. When done, you get a full report with pass/fail results

What's included

• AI QA Engineer agent — creates test plans, executes them, logs bugs with evidence
• Browser testing — headless Chrome with screenshots, DOM inspection, network monitoring
• Test case management — create, run, and track repeatable tests with full history
• Bug tracking — built-in kanban board with severity levels and comments
• Agent orchestration — QA finds bugs, Developer agent fixes them, QA re-tests automatically

No expensive API tokens

QA Panda uses Codex CLI under the hood — it works with your existing ChatGPT subscription. No API keys needed.

When it finds bugs, you can either let the built-in Developer agent fix them, or just copy the report and paste it into Claude Code / Codex / Cursor to fix.

Links

• GitHub: github.com/gzmagyari/qapanda
• VSCode Marketplace: Install QA Panda

Free and MIT licensed. Would love to hear what you think!