Podgen - AI Podcast Generator

Transform your curious questions into natural podcast conversations, and listen to learn from them.

Comment

🎙️ PodGen - AI Podcast Generator

Winter 30 Hackathon Submission

Transform any text content into engaging, natural-sounding podcast conversations with AI-powered script generation and neural text-to-speech.

Demo Video: https://youtu.be/nE_eR3pYIU8

🌟 Features

Content Input Options

  • 📄 File Upload: PDF, TXT, DOCX, HTML, EPUB
  • 🔗 URL Paste: Articles, blogs, Wikipedia pages
  • 🔍 Online Search: AI-powered research using Groq
  • ✏️ Direct Paste: Copy-paste text content

Script Generation

  • 🎭 Two Distinct Characters: Priya (Female Host) & Arjun (Male Co-host)
  • 🗣️ Natural Hinglish: 74% English, 20% Hindi, 3% conversational fillers, 3% formal pauses
  • 💬 Conversational Elements: "Hmmm...", "Acccha...", "Is that so?", "Ohh, I see!"
  • ✨ Professional Tone: Dignified language, no slang

Audio Generation

  • 🔊 Edge TTS Neural Voices: High-quality Microsoft text-to-speech
  • 🇮🇳 Multi-Language Support: English, Hindi, Tamil, Telugu, Bengali, Kannada, Malayalam, Marathi, Gujarati
  • 🎧 Voice Preview: Test voices before generating
  • 📥 Downloadable MP3: Save combined podcast file

Modern UI/UX

  • 🎨 Beautiful Audio Player: Canvas waveform visualization
  • 📝 Editable Names: Rename projects and audio files
  • 💾 Auto-Save: Local storage persistence
  • 📱 Responsive Design: Works on all devices

🚀 Quick Start

Option 1: One-Click Start (Recommended)

Mac/Linux:

chmod +x start.sh
./start.sh

Windows:

start.bat

This will automatically:

  • Install all dependencies
  • Start the backend server (port 8000)
  • Start the frontend server (port 5173)

Option 2: Manual Setup

Backend:

cd backend
python3 -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate
pip install -r requirements.txt

# Create .env file with your Groq API key
echo "GROQ_API_KEY=your_api_key_here" > .env

python main.py

Frontend:

npm install
npm run dev

Option 3: Google Colab (No Setup Required)

  1. Open Podcast_Creator_Colab.ipynb in Google Colab
  2. Run all cells sequentially
  3. Follow interactive prompts

📍 Access Points

🛠️ Technology Stack

Component Technology
Frontend React 18, TypeScript, Tailwind CSS, Vite
Backend FastAPI, Python 3.9+
AI/LLM Groq (Llama 3.1-8B Instant)
TTS Edge TTS (Microsoft Neural Voices)
Audio HTML5 Canvas Visualization

📁 Project Structure

podgen-ai-podcast-generator/
├── src/                          # React frontend source
│   ├── app/
│   │   ├── App.tsx              # Main application
│   │   └── components/
│   │       ├── UploadStep.tsx   # Step 1: Content input
│   │       ├── ScriptStep.tsx   # Step 2: Script generation
│   │       └── AudioStep.tsx    # Step 3: Audio synthesis
│   └── styles/                  # CSS and themes
├── backend/
│   ├── main.py                  # FastAPI server
│   ├── requirements.txt         # Python dependencies
│   └── .env                     # API keys (create this)
├── Winter 30 Hackathon deliverables/
│   ├── backend/                 # Standalone backend
│   ├── frontend/                # Standalone frontend
│   └── docs/                    # Documentation
├── Podcast_Creator_Colab.ipynb  # Google Colab notebook
├── start.sh                     # Mac/Linux startup
├── start.bat                    # Windows startup
├── package.json                 # Node.js config
└── README.md                    # This file

🔧 API Endpoints

Content Processing

  • POST /api/content/wikipedia - Fetch Wikipedia article
  • POST /api/content/perplexity - Search and summarize topics
  • POST /api/content/url - Extract content from URL
  • POST /api/content/upload - Process uploaded files

Script Generation

  • POST /api/script/generate - Generate conversational script
  • POST /api/script/summarize - Summarize long content

Audio Generation

  • POST /api/audio/generate - Generate podcast audio from script
  • GET /audio/{filename} - Serve generated audio files

📖 Usage Guide

Step 1: Content Input

  1. Choose your content source:
    • Wikipedia: Search by topic
    • URL: Paste article link
    • Upload: Select file (PDF, TXT, DOCX, etc.)
    • Search: AI-powered topic research
  2. Click "Upload content from source"
  3. Preview content in the right panel

Step 2: Script Generation

  1. Content is automatically displayed
  2. Click "Generate script"
  3. Review the generated conversation between Priya and Arjun
  4. Click "Next" to proceed

Step 3: Audio Generation

  1. Select voice for Priya (P1)
  2. Select voice for Arjun (P2)
  3. Click "Generate audio"
  4. Listen to the podcast in the player
  5. Download the MP3 file

🎯 Key Innovations

  1. Hinglish Prompting: Carefully engineered prompts for natural Hindi-English code-switching
  2. Character Consistency: Fixed roles (Priya/Arjun) with gender-appropriate voices
  3. Canvas Waveform: Smooth 60fps audio visualization
  4. Smart Naming: Auto-extracts 1-2 keywords from content for project titles
  5. Multi-Source Content: Flexible input from various sources

🔑 API Keys Setup

  1. Get Groq API Key (Required):

  2. Create .env file:

    cd backend
echo "GROQ_API_KEY=your_actual_key_here" > .env

  3. Restart backend if already running

🐛 Troubleshooting

Backend won't start

  • ✅ Ensure Python 3.8+ installed: python3 --version
  • ✅ Check port 8000 available: lsof -i :8000
  • ✅ Verify GROQ_API_KEY in .env file
  • ✅ Install dependencies: pip install -r backend/requirements.txt

Frontend won't start

  • ✅ Ensure Node.js 16+ installed: node --version
  • ✅ Check port 5173 available: lsof -i :5173
  • ✅ Install dependencies: npm install
  • ✅ Clear cache: rm -rf node_modules package-lock.json && npm install

Content fetch fails

  • ✅ Backend must be running on port 8000
  • ✅ Check browser console for errors
  • ✅ Verify GROQ_API_KEY is valid
  • ✅ Check internet connection

Audio generation issues

  • ✅ Edge TTS requires internet connection
  • ✅ Check backend console for errors
  • ✅ Verify script was generated successfully
  • ✅ Ensure audio_output directory exists in backend/

📦 Deliverables

This project includes complete deliverables for the Winter 30 Hackathon:

  • ✅ Full-stack web application (React + FastAPI)
  • ✅ Google Colab notebook for standalone use
  • ✅ Complete documentation and setup instructions
  • ✅ Demo video and usage guide
  • ✅ Source code with detailed comments
  • ✅ Attribution and license information

See Winter 30 Hackathon deliverables/ folder for organized submission files.

📝 Notes

  • Groq API offers generous free tier (14,400 requests/day)
  • Edge TTS is free and requires no API key
  • All generated content saved locally in audio_output/
  • Project state persists in browser local storage
  • Works best with English content for Hinglish output

📄 License

This project is for personal and educational use.

🤝 Contributing

Feel free to submit issues or pull requests!

Built With

Share this project:

Updates