Study Buddy: A Local AI Study Partner Built with Gemma 3

Surya Jyoti Mantry — Sun, 04 Oct 2026 15:08:44 +0000

This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend

What I Built

I built Study Buddy for my friends who need a quick, focused way to prepare for exams at the last minute. They wanted to practise short question-and-answer sessions on different topics and keep important points handy.

Study Buddy lets users save and organize notes, then turn a note into a short summary, five quiz questions, and an answer key. My friends tried it and said it was useful for last-minute exam preparation, practising questions on a topic, and keeping important points in one place.

Demo

https://youtu.be/ddcUgjEjD7s

Code

https://github.com/suryastack/Study-Buddy

How I Built It

The frontend uses React, Redux Toolkit, and CSS. An Express backend handles notes and study-set requests.

For AI, I used Gemma 3 1B, an open-weight model, running locally through Ollama. When someone clicks “Make me a quiz,” the app sends the selected note to Express, which asks the local Ollama instance to generate the study set.

Why Does Open Innovation Matter?

The model runs on my own computer, so notes go to local Ollama instead of a hosted AI API. There’s no AI API key or per-request API charge. After the initial model download, Gemma can generate study sets locally. I can also try a different Ollama-compatible model without rebuilding the whole app.

Prize Categories

Best Use of Gemma