What it does

Questanda took a plain-language question, answered it with Gemini, and stored the question and answer in MongoDB so the answer had a page of its own.

How I built it

Next.js and Tailwind on the front end, the Gemini API for answers, MongoDB for storage, and a generated sitemap so the answer pages could be indexed.

Why I built it

This was my first project on a large language model. I wanted to learn the API end to end, from prompt to stored output to a page a search engine could read.

Where it went

Retired. The domain is offline and the code is still on GitHub. What I learned here about storing and serving model output turned into later work: Mender, which patches other agents' prompts and checks the fix against evals, and Actuary402, which buys from agent services and grades what comes back.

Built With

Share this project:

Updates

Submission history