WebVoyant

Inspiration

We were tired of seeing developers lose sleep over broken websites. You launch a new feature, everything looks perfect, and then users report bugs on Safari or layout issues on mobile. Manual testing across browsers takes hours, so many teams skip it.

We asked: what if testing could be as simple as describing what you want? What if AI could handle “click the login button and check the dashboard” across every browser?

WebVoyant was built to make testing simple and even fun. It doesn’t just find bugs, it celebrates success with confetti.

What it does

WebVoyant is an AI-powered platform that runs end-to-end tests across browsers and devices using ONLY human language prompt. It helps developers automate repetitive testing and catch issues early.

How We Built

We used Next.js, React 19, and TypeScript for the frontend with Tailwind CSS for a clean UI. The backend uses FastAPI with Python, Playwright, powered by the Gemini 2.5 model which handles cross-browser automation on Chrome, Firefox, and Safari. We also added analytics, geographic data, and visual success effects.

Accomplishments that we're proud of

We integrated Gemini for smart test generation and Statsig for analytics visualization, turning raw automation data into insights developers actually enjoy using.

Challenges we ran into

Our biggest challenge was deploying Statsig onto our pipeline and integrating it seamlessly with our existing automation workflow. We struggled with API compatibility issues between our FastAPI backend and Statsig's JavaScript SDK, requiring extensive middleware development to bridge the gap between Python and JavaScript environments.

What we learned

Through our journey of building WebVoyant, we learned how to develop a user-friendly UI, how to deploy different API keys and authorize, and how to debug a complex system.