Urgent.News

What's breaking now, across thousands of outlets.

AI

Exam Buddy: A Local Gemma 3 Quiz Tool I Built for a Friend's OS Exam

This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend What I Built A friend of mine was preparing for an Operating Systems exam , and his study method was the one most of us fall back on: read the notes again, and again, and feel like it's working. Re-reading feels productive, but it doesn't tell you what you actually know. A quiz does. So I built Exam Buddy , a study…

This story is about a local AI-powered quiz tool called Exam Buddy, built by the author to help a friend prepare for an Operating Systems exam. The tool works by taking personal notes and turning them into quiz questions, ensuring that the questions reflect the material covered in the notes.

Exam Buddy offers three quiz modes: multiple-choice, written answers graded by the AI, and a mixed mode. Users can choose the number of questions per quiz, which ranges from 5 to 15. The app also covers the entire notes file by splitting long notes into parts, one part per quiz.

The author explains that the tool provides personalized explanations in a theme chosen by the user, ensuring that even wrong answers are accompanied by explanations that make sense for the user. For instance, a wrong answer about process scheduling can be explained using football tactics, or a kitchen scenario with too few cooks.

The tool also includes a follow-up Q&A mode, providing additional quizzes only for the questions the user got wrong. The score history feature tracks the user's performance over time, and a "what to review" summary is generated after each quiz.

The entire operation of Exam Buddy is fully local, meaning that notes never leave the user's machine. To run the tool, users need to install Ollama, pull Gemma 3 4B, and serve the web UI with Python's http.server module. The code is available on GitHub (Phantom9869 / exam-buddy).

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in AI

想知道 Claude Code 額度被哪個專案吃掉,就寫了個小工具

撞到 Claude Code 5 小時或每週上限的時候,我最想知道的是哪個專案用掉的,但 /usage 答不出來。它只給帳號層級的百分比,再加幾個像「>150k context」、「4+ parallel sessions」的標籤,沒有拆到專案,那些百分比也沒辦法自己驗算。 Claude Code 本來就會把每則訊息的 token 用量寫進本機的 session transcript(…

  • Reporter creates 'quota-audit' tool to track Claude Code usage by project.
  • Tool reads session transcripts from ~/.claude/projects//.jsonl files.
  • In 7-day window, 65% of $147.31 cost allocated to 'my-app' project.

More from Monday 5 October →