PitchPulse - AI-Powered Presentation Practice Tool

PitchPulse is an interactive web application that helps users practice and improve their presentation skills using AI-powered feedback. The tool tracks eye contact, speaking pace, and provides real-time feedback to enhance presentation confidence.

Features

• Real-time Eye Contact Tracking: Monitors and provides feedback on eye contact percentage during practice sessions
• Speaking Pace Analysis: Tracks words per minute and provides feedback on speaking speed
• Confidence Scoring: Calculates an overall confidence score based on eye contact and speaking metrics
• Multiple Difficulty Levels: Choose from Beginner, Intermediate, and Advanced practice sessions
• Interactive Feedback: Real-time visual feedback on performance metrics
• Session Scoring: End-of-session scoring system (0-100) based on:
  • Eye contact percentage
  • Speaking pace
  • Overall confidence
  • Session completion

Technical Stack

• React + TypeScript
• Vite
• Tailwind CSS + DaisyUI
• TensorFlow.js (for face detection)
• Web Speech API (for speech recognition)
• WebRTC (for camera access)

Prerequisites

• Node.js (v16 or higher)
• Modern web browser with camera access
• WebRTC support
• Web Speech API support

Installation

1. Clone the repository:

git clone https://github.com/Dimural/PitchPulse.git
cd PitchPulse

2. Install dependencies:

npm install

3. Download required AI models:

npm run download-models

4. Start the development server:

npm run dev

5. Open your browser and navigate to the URL shown in your terminal (typically http://localhost:3000 or higher)

Usage

1. Select a difficulty level (Beginner, Intermediate, or Advanced)
2. Click "Start Session" to begin
3. Speak naturally while maintaining eye contact with the camera
4. View real-time feedback on your performance
5. End the session to see your final score and detailed feedback

Performance Metrics

• Eye Contact: Percentage of time maintaining eye contact with the camera
• Speaking Pace: Words per minute (WPM) calculation
• Confidence Score: Combined metric based on eye contact and speaking pace
• Final Score: Overall score (0-100) considering all metrics and session completion

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.