Ellipsis | One-Click Podcast Generation Agent

A next-gen podcast generation agent that brings human-like, high-quality audio content to life on any topic with just one click

Ellipsis is a next-generation podcast generation agent that brings human-like, high-quality audio content to life on any topic with just one click. Whether it’s breaking news, deep-dive tech explainers, movie reviews, or post-match sports breakdowns, Ellipsis crafts intelligent podcast episodes that sound like they were created by seasoned hosts in a professional studio.

Features

• Intelligent Multi-Speaker Dialogue with multiple distinct voices and personalities
• Comprehensive Topic Coverage from LLM architectures to lunar eclipses
• Custom Evaluation Engine ensuring factual accuracy, legal compliance, and conversational quality
• Fully Automated Podcast Generation with human-like, podcast-ready audio output
• Real-time Streaming Updates via Server-Sent Events (SSE)
• Podbean Integration for direct podcast publishing
• Trending Topics Detection using Perplexity API

Prerequisites

• Node.js v16+ and npm/yarn
• Python 3.10+ and pip
• Redis server running (default on port 6380)
• Perplexity API key, Podbean credentials

Installation

# Clone the repository
git clone https://github.com/dineshkannan010/Ellipsis.git
cd Ellipsis

# Backend setup
cd backend
python -m venv venv
source venv/bin/activate    # macOS/Linux
pip install -r requirements.txt

# Install native packages
pip install llama-cpp-python --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cpu
pip install git+https://github.com/freddyaboulton/orpheus-cpp.git
pip install huggingface_hub[hf_xet] hf_xet

# Frontend setup
cd ../frontend
npm install

Configuration

Create backend/.env:

REDIS_URL=redis://your-redis-host:6379
PERPLEXITY_API_KEY=your_key_here
PODBEAN_CLIENT_ID=...
PODBEAN_CLIENT_SECRET=...

Create frontend/.env:

REACT_APP_API_URL=http://your-backend-host:5000

Usage

1. Start Redis Server:

   redis-server --port 6380
   

2. Launch Backend:

   cd backend
   python app.py
   

3. Launch Frontend:

   cd frontend
   npm run dev
   

4. Optional: Podbean Integration:

   cd backend/integrations/podbean_mcp
   pip install -e .
   python server.py
   python client.py server.py
   

5. Generate Content: Enter a topic in the homepage textbox and hit Enter. Switch to ContentGenerationView to see live script & audio progress.

Code Explanation

• Backend: Python Flask with Redis pub/sub, llama.cpp, and Orpheus TTS for audio generation
• Frontend: React with Vite, Tailwind CSS, and Server-Sent Events for real-time updates
• AI Integration: Perplexity API for content generation and trending topics detection
• Audio Processing: Multi-speaker TTS with distinct voice personalities
• Content Evaluation: Built-in pipelines for factual accuracy and legal compliance
• Podcast Publishing: Direct integration with Podbean via MCP server

Demo Video

Links

• GitHub Repository
• Devpost Submission