Project Info
Inspiration
Quizcribe is inspired by the problem that many students face in retaining the knowledge that they learned from class lectures or open-course videos. Through AI-powered summaries and interactive quizzes, we hope to make learning more effective and expand the effectiveness of self-regulated learning.
What it does
Quizcribe is an educational software that supplements students with AI-powered study resources generated from inputted videos or audio. By entering a link to a video, such as one from YouTube, Quizcribe automatically transcribes the video, returns a detailed summary of its content, and generates an interactive quiz game for knowledge testing and review.
How we built it
Frontend: Next.js/React Backend: Django/Python APIs: Deepgram (Speech to Text), Google Gemini 1.5, YouTube DL
Challenges we ran into
Extracting video URL from web URL (Resolve) Gemini prompt design and output format had to be very precise (Resolved) Frontend and backend connection issues with generated interactive quizzes (Resolved)
Accomplishments we're proud of
The summaries and interactive quizzes generated by Quizcribe is very accurate to the content of the inputted video/audio Our team was able to extend Quizcribe's video transcription and processing service to 16 different languages
What we learned
Fast-paced app development Use of AI and LLM in video/audio transcription and processing Using Next.js for frontend development Developing REST API in Django Version control with Git
What's next
Support for video file upload and other URL domains e.g. Coursera, Khan Academy, Zoom Enable translation to languages other than English Expanding support for more than 16 languages Increasing the efficiency and accuracy of video transcription and summary/quiz generation
Quizcribe
Calhacks 11.0 Michael Siu, Richard Peng, Yiling Yuan, Zhen Liu
Overview
Quizcribe is an educational software that supplements students with AI-powered study resources generated from inputted videos or audio. By entering a link to a video, such as one from YouTube, Quizcribe automatically transcribes the video, returns a detailed summary of its content, and generates an interactive quiz game for knowledge testing and review.
Through providing a more versatile, time-efficient, and fun study medium, Quizcribe aims to expand the effectiveness of free online education, further expanding the accessibility of educational resources to people of all backgrounds.
Main Features:
- Automated speech-to-text transcription that converts videos/audio of any length into text.
- AI-powered video summary and interactive quiz generation.
- Video detection and transcription support for 16 languages.
Running & Installation
To install dependencies:
pip install -r requirements.txt
cd calhacksmryz
npm install
Run App and API Server:
npm run dev
cd backend
python manage.py runserver
The app should be at localhost:3000
Tech Stack
- Frontend: Next.js/React, Tailwind CSS, Material UI
- Backend: Django/Python
- API: Deepgram (Speech to Text), Google Gemini 1.5, YouTube DL
Key Functionalities
Video/Audio to Transcript Generation using Deepgram
- Users are prompted to input a link, which is passed to the YouTube DL and extracted into a video file URL.
- Deepgram AI's Speech-To-Text API is then used to convert the audio into transcripts.
- Deepgram AI's built-in language detection system allows for the transcription of videos/audio in 16 languages.
Content Summarization using Google Gemini
- Gemini API is used to automatically provide consistently structured summaries for the plain text transcripts.
Interactive Quiz Generation using Google Gemini
- Using a chain of thought approach, problem sets are generated from the summarized content. The logic behind this feature involves step-by-step prompt engineering that ensures the generated questions are relevant, thought-provoking, and aligned with the key concepts in the summarized content.
- Corrent answer are determined by AI-powered logic reasoning.
Demo
Summary to video url provided
Quiz
Result
Analysis
View
Metric
- 71
- 33
- 32
- 20
Figures cover GitHub contributors during the hackathon window. A co-authored commit counts in full for each author, so per-member totals add up to more than the whole-team figures.
Technology
- CSSIn code
- DjangoIn code
- JavaScriptIn code
- Next.jsIn code
- PythonIn code
- ReactIn code
- Tailwind CSSIn code
- Google GeminiClaimed
7 of 8 appear in the indexed code. 1 claimed on Devpost could not be matched to code, which may simply mean the tool leaves no trace in the repository.
AI coding agents
No AI coding agent signals were found in this repository.
Detected from committed agent config files and commit authorship. Absence of a signal is not proof an agent was unused.
Codebase size
Source size
164 KB
Source files
25
Counts recognized source files only; vendored directories, binaries and lockfiles are excluded, so this is smaller than the repository on disk.
Repository
notrichardpeng/Quizcribe
37 files · 535 KB · @ 0307c9f
Structure
Interface
6 files · 16%Screens, components and styles rendered to the user.
API & routing
1 file · 3%Request entry points: routes, handlers and controllers.
Application logic
14 files · 38%Domain rules, services and shared utilities.
Supporting
Layers are inferred from where files sit in the tree, not from reading the code. A project that names its directories unconventionally will read oddly here — open the file browser to check anything the diagram implies.
Languages
- YAML77%
- JavaScript11%
- Python9%
- Markdown2%
- CSS0%
Share of indexed source by file size. Binary and vendored files are excluded.
Dependencies
calhacksmryz/package.json
npm · 12- @emotion/react
- @emotion/styled
- @mui/icons-material
- @mui/material
- next
- react
- react-dom
- react-simple-typewriter
- +4 more
requirements.txt
pypi · 7- deepgram_sdk
- Django
- django-cors-headers
- djangorestframework
- google-generativeai
- python-dotenv
- yt_dlp
Declared in the repository’s manifests at the indexed commit. A declared package is not proof it is used, and runtime dependencies are listed first.
This project’s features have not been analysed yet.
Export this project's context (description, README, evidence, key source files) to chat with an AI agent elsewhere.