Lec-Recall

Heidi Health for Lectures - making lectures interactive

Comment

🎓 Lec-Recall

Your intelligent learning companion with AI-powered question detection and real-time student engagement

Lec-Recall transforms traditional lectures into interactive, engaging experiences by automatically detecting questions in speech, generating real-time quizzes, and providing comprehensive analytics for both lecturers and students.

✨ Features

For Lecturers

  • 🎤 Real-time Speech Transcription - Powered by Deepgram API for accurate voice-to-text conversion
  • 🤖 AI Question Detection - Google Gemini AI automatically identifies questions in your lecture
  • ⚡ Instant Quiz Generation - Convert detected questions into multiple-choice quizzes automatically
  • 📊 Live Analytics Dashboard - Monitor student engagement, quiz performance, and confusion levels
  • 🤔 Confusion Meter - Real-time feedback on student understanding levels
  • 📈 Session Analytics - Detailed post-session reports with performance insights
  • 🎯 Review Recommendations - Identifies topics that need reinforcement based on student performance

For Students

  • 🔗 Easy Session Joining - Join lectures with simple 6-character codes
  • 📝 Interactive Quizzes - Participate in real-time quizzes generated from lecture content
  • 🤔 Anonymous Feedback - Signal confusion levels without revealing identity
  • 📊 Personal Analytics - View your performance, correct/incorrect answers, and improvement areas
  • 📚 Lecture Summaries - AI-generated summaries of lecture content for review
  • 🌙 Dark Mode Support - Comfortable viewing in any lighting condition
  • ♿ Full Accessibility - WCAG 2.1 AA compliant with screen reader support

🏗️ Architecture

The application follows a modern client-server architecture with real-time communication:

Technology Stack

  • Frontend: React 19, Socket.IO Client, CSS3 with custom properties
  • Backend: Node.js, Express.js, Socket.IO Server
  • Database: SQLite with structured schema
  • AI Services: Google Gemini AI for content processing
  • Speech Recognition: Deepgram API for real-time transcription
  • Real-time Communication: WebSocket connections via Socket.IO

🔄 How It Works

  1. Lecturer Setup

    • Create a session with a name and time limit
    • Start recording your lecture using your microphone
    • The system automatically detects questions you ask
    • Students join using a simple 6-character code

  2. Real-Time Interaction

    • Questions are converted into multiple-choice quizzes instantly
    • Students answer on their devices in real-time
    • Monitor student understanding through the confusion meter
    • See live analytics of student performance

  3. Session Completion

    • End the session to generate comprehensive analytics
    • Review which topics need reinforcement
    • Students receive personalized performance reports
    • AI-generated lecture summaries help with review

Data Flow

  1. Speech Processing: Deepgram API converts lecturer audio to text
  2. AI Analysis: Gemini AI processes transcripts to detect questions
  3. Quiz Generation: Detected questions are converted to multiple-choice format
  4. Real-time Distribution: Socket.IO broadcasts quizzes to connected students
  5. Response Collection: Student answers are stored and analyzed in real-time
  6. Analytics Generation: Performance metrics and insights are calculated

🎨 Design Features

Modern UI/UX

  • Glass morphism design with backdrop blur effects
  • Responsive layout that works on desktop, tablet, and mobile
  • Dark/light mode with system preference detection
  • Smooth animations and micro-interactions
  • Color-coded feedback for different states and actions

Accessibility (WCAG 2.1 AA Compliant)

  • Screen reader support with proper ARIA labels
  • Keyboard navigation for all interactive elements
  • High contrast ratios for text readability
  • Skip links for efficient navigation
  • Semantic HTML structure
  • Focus indicators for better usability
Share this project:

Updates