🚀 IntelliVoice – AI Voice Coaching & Real-Time Mock Interview Platform

Your Personal AI Coach for Interviews, Lectures, GD, and Communication Skills.

IntelliVoice is an advanced AI-powered voice coaching platform built using Next.js, Convex, Deepgram, and GPT models. It analyzes your speech in real-time, transcribes your audio, and generates expert-level responses based on selected coaching modes like:

• 🎤 Mock Interview Expert
• 📘 Lecture Mode
• 🎯 GD / Debate Mode
• 🧠 Career & Skill Coach
• 💼 HR Interview Mode

IntelliVoice helps users practice speaking, improve fluency, and provides AI feedback—making it perfect for interview prep, communication training, and learning.

---

🌟 Key Features

🎙️ Real-Time Voice Transcription

• Powered by Deepgram Streaming APIs
• High accuracy + low latency

🧠 AI Response Generation

• Uses GPT-based model
• Custom expert prompts depending on selected mode (Interview, Lecture, GD)

🚦 Mode-Based Intelligence

Each mode has unique behavior:

• Interview Mode: Asks follow-up questions, gives feedback
• Lecture Mode: Explains concepts with depth
• GD Mode: Argues both sides logically
• Coach Mode: Encourages improvement and gives tips

🔐 Custom Authentication System (Integrated with Convex)

• Fully custom-built authentication system, no external auth providers

• Uses Convex database to store users, hashed passwords, and sessions

• Implements secure password hashing (bcrypt/argon2)

• Generates JWT-based tokens for sessions

• Convex functions handle:

  • User registration
  • Login validation
  • Token verification
  • Protected server routes

• Frontend uses tokens to authenticate all API calls

⚡ Real-Time Backend (Convex)

• Stores users, sessions, transcripts, audio logs
• Fast and reactive

🧩 Modern UI/UX

• Built with Next.js App Router + Tailwind CSS + ShadCN UI
• Smooth animations
• Responsive design

🗂️ Session Dashboard

• Tracks your progress
• Shows session history
• Quick launch for coaching modes

---

🛠️ Tech Stack

Frontend

• Next.js 14
• React
• Tailwind CSS
• ShadCN UI
• Framer Motion

Backend

• Convex (real-time database + API)
• Deepgram (audio streaming)
• OpenAI GPT
• Custom Authentication System (JWT + hashed passwords)**

---

📁 Project Structure

intellivoice/
 ├── app/
 │   ├── dashboard/
 │   ├── discussion/
 │   └── api/
 ├── components/
 │   ├── VoiceChat.tsx
 │   ├── ExpertList.ts
 │   ├── CoachingExpert.ts
 ├── convex/
 │   ├── functions/
 │   └── schema.ts
 ├── lib/
 ├── public/
 └── README.md

---

🚀 Getting Started

1️⃣ Clone the repo

git clone https://github.com/your-username/intellivoice.git
cd intellivoice

2️⃣ Install dependencies

npm install

3️⃣ Add environment variables

Create .env.local:

AUTH_SECRET_KEY=
AUTH_JWT_KEY=

NEXT_PUBLIC_DEEPGRAM_KEY=
OPENAI_API_KEY=

CONVEX_DEPLOYMENT=

4️⃣ Setup Convex

npx convex dev

5️⃣ Run project

npm run dev

---

🎥 How It Works (Flow)

1. User selects coaching mode (Lecture, Interview, GD…)
2. Fills session form
3. Starts speaking → Deepgram streams transcription
4. Transcription triggers AI prompt
5. AI sends real-time response
6. User sees & hears the reply
7. Entire session saved in Convex backend

---

🤝 Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you'd like to change.

---

⭐ Support & Feedback

If you like this project, consider giving it a ⭐ on GitHub! Your support motivates development!

---