Voicebook - PDF & YouTube to Podcast Converter

Voicebook is a sophisticated web application that transforms PDF documents and YouTube videos into engaging podcast-style conversations between two AI personas (Alex and Sarah). The application leverages advanced AI technologies for content processing, embedding generation, and natural language understanding.

🚀 Features

PDF Processing

• Upload and process PDF documents
• Automatic content chunking and embedding
• Vector storage in Pinecone database
• Generate AI-powered podcast conversations
• Interactive chat with document context

YouTube Integration

• Process YouTube videos via URL
• Automatic caption/transcript extraction
• Content vectorization and storage
• Generate podcast discussions about video content
• Context-aware chat about video content

Podcast Generation

• Dynamic conversation generation between Alex and Sarah
• Minimum 55 lines of detailed dialogue
• Structured discussion format:
  • Introduction/Overview
  • Main points analysis
  • Critical discussion
  • Real-world implications
  • Personal perspectives

Interactive Features

• Real-time audio playback
• Voice-enabled chat interface
• Context-aware responses
• PDF/Video content reference
• Semantic search capabilities

🛠 Tech Stack

Frontend

• Next.js 15.0.3 - React framework
• React 19 - UI library
• TailwindCSS - Styling
• TypeScript - Type safety
• React Icons - Icon components

Backend (API Routes)

• Next.js API Routes - Serverless functions
• Pinecone - Vector database
• Google AI (Gemini) - Embeddings generation
• Groq - LLM for conversation generation
• LangChain - Document processing
• PDF Parse - PDF text extraction

AI/ML Components

• Gemini Embedding Model - Vector embeddings
• Llama 3.2 90B - Podcast generation
• Mixtral 8x7B - Chat responses
• Web Speech API - Voice interface

📦 Key Dependencies

{
  "@google/generative-ai": "^0.21.0",
  "@langchain/community": "^0.3.14",
  "@pinecone-database/pinecone": "^4.0.0",
  "groq-sdk": "^0.8.0",
  "langchain": "^0.3.5",
  "next": "15.0.3"
}

🏗 Architecture

Document Processing Flow

1. PDF/YouTube content upload
2. Content chunking and preprocessing
3. Embedding generation via Gemini AI
4. Vector storage in Pinecone
5. Podcast script generation via Groq
6. Interactive chat capabilities

Data Flow

1. Content Ingestion → Chunking → Embedding → Storage
2. Query Processing → Semantic Search → Context Retrieval → Response Generation
3. Chat Interface → Voice Processing → Context-Aware Responses

🔧 Environment Setup

1. API Keys Required

Pinecone API Key

1. Visit Pinecone Console
2. Sign up or login to your account
3. Navigate to API Keys section
4. Create a new API key
5. Copy the key and environment

Google AI (Gemini) API Key

1. Go to Google AI Studio
2. Create or sign in to your Google Cloud account
3. Enable the Gemini API
4. Create a new API key
5. Copy the key

Groq API Key

1. Visit Groq Console
2. Create an account or sign in
3. Go to API section
4. Generate new API key
5. Copy the key

2. Environment Configuration

1. Clone the repository:

git clone https://github.com/yourusername/voicebook.git
cd voicebook

2. Copy the environment example file:

cp .env.example .env

3. Update the .env file with your API keys:

PINECONE_API_KEY=your_pinecone_api_key
PINECONE_INDEX_NAME=your_index_name
PINECONE_ENVIRONMENT=your_environment
GOOGLE_API_KEY=your_google_api_key
GROQ_API_KEY=your_groq_api_key

3. Pinecone Index Setup

1. Create a new index in Pinecone console with:

   • Dimensions: 768 (Gemini embeddings)
   • Metric: Cosine
   • Pod Type: s1.x1 (recommended)

2. Update your .env with the index name:

PINECONE_INDEX_NAME=your-index-name

4. Development Setup

1. Install dependencies:

npm install
# or
yarn install

2. Run the development server:

npm run dev
# or
yarn dev

3. Build for production:

npm run build
npm start
# or
yarn build
yarn start


## 🎯 Key Components

### Content Processing
- PDF document chunking and embedding generation
- YouTube transcript extraction and processing
- Vector storage and retrieval

### Conversation Generation
- Structured podcast script generation
- Context-aware chat responses
- Voice interface integration

### User Interface
- Responsive design with TailwindCSS
- Audio playback controls
- Interactive chat interface
- Voice command support

## 📝 API Routes

### Main Endpoints
- `/api/upload` - PDF processing
- `/api/youtube` - YouTube video processing
- `/api/generate-podcast` - Podcast script generation
- `/api/chat` - Context-aware chat
- `/api/podcast-chat` - Podcast-specific chat
- `/api/podcast-yt-chat` - YouTube podcast chat

## 🔒 Security Considerations

- Environment variables for API keys
- Server-side processing of sensitive operations
- Rate limiting implementation
- Error handling and validation

## 🎨 UI/UX Features

- Clean, modern interface
- Responsive design
- Loading states and animations
- Error handling and user feedback
- Voice interaction capabilities

## 📚 Documentation References

- [Next.js Documentation](https://nextjs.org/docs)
- [Pinecone Documentation](https://docs.pinecone.io/)
- [Google AI Documentation](https://ai.google.dev/docs)
- [Groq Documentation](https://console.groq.com/docs)
- [LangChain Documentation](https://js.langchain.com/docs)

## 🤝 Contributing

Contributions are welcome! Please read our contributing guidelines and submit pull requests for any enhancements.

## 📄 License

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