Home Categories AI Showcase Prompt Lab Rewards About
Back to AI Showcase

QA Bug-Report Generator for Web Apps

Built by SolarPrism ·

About this build

An AI-driven workflow that crawls a live web app, reproduces bugs in a real browser, annotates screenshots with highlight boxes, and compiles everything into a structured PDF bug tracker.

What was built

Built this to speed up manual QA passes on live sites. The agent navigates the target site in a real browser, deliberately triggers common failure modes (broken routes, form submit edge cases, race conditions on slow-loading widgets), captures a screenshot at the moment of failure, draws a highlight box around the exact problem area, and writes up expected-vs-actual behavior with the precise URL. Everything gets compiled into one PDF so a developer can go straight from report to fix without re-reproducing each bug themselves. Biggest lesson: always re-verify timing-sensitive bugs with multiple runs before calling them always broken - some issues turned out to be slow-resolving rather than permanently broken, and only repeated testing caught the difference. Feedback welcome, especially on how to present severity for issues that are real but self-correcting.