https://github.com/ryu-ryuk/exam-whisperer

a smart ai-powered exam revision assistant that explains concepts, quizzes you, and tracks your topic mastery in real time with a gamified leaderboard system
https://github.com/ryu-ryuk/exam-whisperer

Last synced: 8 months ago
JSON representation

a smart ai-powered exam revision assistant that explains concepts, quizzes you, and tracks your topic mastery in real time with a gamified leaderboard system

• Host: GitHub
• URL: https://github.com/ryu-ryuk/exam-whisperer
• Owner: ryu-ryuk
• License: mit
• Created: 2025-06-27T12:17:03.000Z (12 months ago)
• Default Branch: main
• Last Pushed: 2025-07-05T04:01:34.000Z (12 months ago)
• Last Synced: 2025-07-05T05:18:50.049Z (12 months ago)
• Language: TypeScript
• Homepage: https://whisper-rust.vercel.app
• Size: 3.01 MB
• Stars: 0
• Watchers: 0
• Forks: 1
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          


  

  


  Whisper





  a real-time, ai-powered exam revision assistant.





  

    

  

  

    

  

  

    

  





  

  

  

  





  Whisper is a real-time exam revision agent that helps you learn, not just get answers.


  it explains concepts, quizzes you, and tracks topic mastery using LLMs and Pathway in a continuous feedback loop.



  not a chatbot, but a learning engine.



---

## features

- explain any concept in simple language  

- generate topic-specific quizzes (MCQ, short answers)  

- track your scores and mastery per topic  

- real-time adaptive learning using Pathway  

- switchable model support: OpenAI, Gemini, Ollama, local models  

- beautiful and responsive UI (Next.js + Tailwind)  

- user auth with email/password or OAuth login  

- fully documented API in [`docs/api.md`](docs/api.md)

---

## how it helps you learn

traditional llm tools give you answers.  

whisper helps you build understanding.

- monitors your progress by topic  

- adapts questions based on performance  

- suggests revisions when needed  

- real-time feedback loop (no cron jobs or polling)

---

## architecture

**core backend**

- `fastapi` — api framework  

- `pathway` — stream computation (topic mastery tracking)  

- `postgres` — persistent store  

- `ollama`, `openai`, `gemini` — interchangeable llm backends  

**frontend**

- `next.js` with `tailwindcss`  

- login + protected routes  

- clean, minimal ux focused on learning flow

---

## quickstart

```bash

# from backend/

make rebuild

make up

# run adaptive learning flow

python3 src/pathway_flow/pathway_main.py