1. LectureNotes AI: Voice-to-Notes Generator

An AI-powered tool that converts audio lectures into structured notes, summaries, and quizzes in seconds

2. Short Description/Purpose

This project solves a core student problem: it's nearly impossible to listen actively and take detailed notes simultaneously. LectureNotes AI automates this by:

• 🎧 Converting spoken lecture audio into accurate text
• 📝 Summarizing key points into clear, study-ready notes
• ❓ Generating interactive quizzes to test understanding

The goal is to help students capture every important concept without the stress of missing information during live lectures.

3. Tech Stack

Component	Technology Used
Backend Framework	Python, Flask
Speech-to-Text Engine	SpeechRecognition, Google Web Speech API
AI Processing Logic	Custom rule-based summarization & quiz generation
Frontend	HTML5, CSS3, Vanilla JavaScript
Audio Processing	Pydub, FFmpeg
UI/UX Design	Custom CSS with ocean-themed gradients & animations
Deployment	Local server (Flask)

4. Data Flow & Processing

Input: Audio files (MP3, WAV, M4A) from student lectures
Step 1 – Upload: User drags/drops file into browser (max 150MB)
Step 2 – Transcription: Backend converts audio to text using speech recognition
Step 3 – AI Analysis:

• Summarization: Extracts key sentences based on importance scoring
• Quiz Generation: Creates fill-in-the-blank questions from key terms
  Step 4 – Output: Three-panel dashboard with Transcript, Summary, and Quiz

5. Features/Highlights

📊 Key Metrics Displayed

✅ Full Transcript – Complete text conversion of the lecture
✅ AI Summary – Concise overview of main points (rule-based/transformer-based)
✅ Interactive Quiz – 10-15 fill-in-the-blank questions per lecture
✅ Real-time Processing – Visual progress bar during audio analysis

⚙️ Technical Capabilities

🔊 Audio Support: MP3, WAV, M4A, FLAC formats
🧠 Smart Blank Selection: Avoids common words, selects meaningful terms for quizzes
📱 Responsive UI: Works on desktop & mobile with ocean-themed design
📈 Progress Tracking: Step-by-step feedback during processing

👨‍🎓 User Workflow

1. Upload lecture audio via drag-and-drop
2. Watch real-time processing progress
3. Access three organized panels:
   • 📄 Transcript for full reference
   • 📝 Summary for quick review
   • ❓ Quiz for self-testing
4. Toggle quiz answers with click interaction

🚀 Performance Features

⚡ Chunked Processing: Handles long lectures by splitting audio
⚡ Optimized Conversion: FFmpeg ensures fast MP3-to-WAV conversion
⚡ Smart Question Generation: Scores sentences for quiz-worthiness
⚡ Uniform Layout: All output panels have consistent, scrollable design

6. Sample Input/Output

Input:

18-minute Computer Science lecture on "Machine Learning Basics" (MP3)

Output:

Transcript:

2,800-word accurate text of the entire lecture

Summary:

"Machine learning enables computers to learn from data without explicit programming. Three main types exist: supervised (labeled data), unsupervised (pattern finding), and reinforcement (reward-based) learning. Neural networks mimic the brain's structure for deep learning tasks."

Quiz Questions:

1. "______ learning uses labeled data to train models for predictions."
   Answer: Supervised
2. "Neural networks are inspired by the structure of the human ______."
   Answer: brain
3. "When a model performs well on training data but poorly on new data, it's called ______."
   Answer: overfitting

7. Setup & Installation

Prerequisites:

• Python 3.8+
• FFmpeg installed and added to PATH

Steps:

1. Clone repository:

   git clone https://github.com/ThisAkshat/lecture-notes-ai.git
   cd lecture-notes-ai

2. Install dependencies:

   pip install -r requirements.txt

3. Run the application:

   python app.py

4. Open browser to: http://localhost:10000

8. Screenshots: