SpeechAssist is an AI-powered interview practice tool designed to help high school students prepare for interviews by providing feedback, transcription, and video analysis in a user-friendly platform. Built during YRHacks 2024, the tool leverages OpenAI Whisper, GPT-4, and video recording technologies to simulate a personalized interview experience.

Interviews can be intimidating, especially for students preparing for university or college admissions. We wanted to create an interactive tool that not only helps students practice but also provides actionable feedback to boost their confidence.

• 🎥 Video Recording: Students record their interview sessions directly in the app.
• 📝 Speech Transcription: Transcribes spoken responses into text using OpenAI Whisper.
• 🔍 AI-Powered Feedback: Utilizes GPT-4 to provide constructive feedback on tone, content, and delivery.
• 🎯 Realistic Questions: Access to a curated bank of common university and college interview questions.
• 📊 Progress Tracking: Track improvement over multiple practice sessions.

• Frontend: React
• Backend: Node.js with Express
• AI Integration: OpenAI Whisper for transcription and GPT-4 for feedback
• Video Handling: WebRTC for seamless video recording
• Deployment: Deployed on Vercel for fast and scalable hosting

1. 🎤 Record: Users record answers to pre-selected interview questions.
2. 📝 Transcribe: Responses are automatically transcribed into text.
3. 🤖 Analyze: AI analyzes the response, providing detailed feedback on structure, clarity, and delivery.
4. 📈 Improve: Users can review feedback and iterate on their responses.

• Pivoting Midway: Initially started as a writing tool but realized it didn’t fit the hackathon submission criteria.
• Speech-to-Text Accuracy: Ensuring transcription quality with noisy inputs.
• Real-Time Feedback: Designing concise yet actionable feedback that resonates with users.

• Fully functional prototype developed within 36 hours!
• Smooth integration of advanced AI tools with a React-based frontend.
• Positive reception from hackathon judges, leading us to be finalists.

• 🎨 Improved UI: Design a more intuitive and user-friendly interface.
• 💬 Expanded Question Bank: Add more diverse and challenging questions.
• 🌍 Accessibility: Optimize for non-native English speakers.
• 📱 Mobile Support: Build a mobile app for on-the-go practice.

• React 🌐
• Node.js ⚡
• OpenAI Whisper 🎙️
• GPT-4 🤖
• WebRTC 🎥

🔗 SpeechAssist on Devpost

We’d love your feedback! Feel free to reach out or contribute to this project. Let’s make interviews less scary for everyone! 💪✨