🎙️ Blog2Podcasts

An AI-powered tool that converts any blog article into a podcast audio file with optional voice cloning from YouTube.

Features

• 🌐 Web Scraping: Extracts main content from any blog URL using trafilatura
• 🤖 AI Summarization: Converts articles into engaging podcast scripts using local LLMs (Ollama)
• 🎵 Text-to-Speech: Generates high-quality audio using Microsoft Edge TTS (free)
• 🎤 Voice Cloning: Clone voices from YouTube videos using Coqui TTS (XTTS-v2)

Installation

From PyPI

pip install blog2podcasts

From Source

git clone https://github.com/QuantBender/blog2podcasts.git
cd blog2podcasts
pip install -e .

With Voice Cloning Support

pip install blog2podcasts[voice-cloning]

Prerequisites

1. Install Ollama

# Linux
curl -fsSL https://ollama.ai/install.sh | sh

# macOS
brew install ollama

# Start Ollama service
ollama serve

2. Pull an LLM Model

# Recommended: Llama 3.2 (fast and capable)
ollama pull llama3.2

# Alternative options:
ollama pull mistral      # Fast, good quality
ollama pull llama3.1     # More capable, slower
ollama pull phi3         # Small, fast

3. Install ffmpeg (for audio processing)

# Ubuntu/Debian
sudo apt install ffmpeg

# macOS
brew install ffmpeg

# Windows
# Download from https://ffmpeg.org/download.html

Usage

Command Line

# Convert a blog to podcast
blog2podcasts https://example.com/blog-article

# Use a different voice
blog2podcasts https://example.com/blog --voice en-GB-RyanNeural

# Use a different LLM model
blog2podcasts https://example.com/blog --model mistral

# Adjust script length (words)
blog2podcasts https://example.com/blog --length 1200

# Preview script without generating audio
blog2podcasts https://example.com/blog --preview

# List available voices
blog2podcasts --list-voices

# Custom output name
blog2podcasts https://example.com/blog -o my_podcast

# Adjust speech rate
blog2podcasts https://example.com/blog --rate "+10%"

🎤 Voice Cloning from YouTube

Clone any voice from YouTube videos and use it for your podcasts!

# Clone voice from a YouTube video
blog2podcasts --clone-voice "https://www.youtube.com/watch?v=VIDEO_ID" --voice-name "my_host"

# Generate podcast with cloned voice
blog2podcasts https://example.com/blog --use-cloned-voice my_host

Python API

from blog2podcasts import BlogScraper, ContentSummarizer, AudioGenerator
from blog2podcasts.cli import BlogToPodcastAgent, PodcastConfig

# Create agent with custom config
config = PodcastConfig(
    voice="en-US-JennyNeural",  # Female US voice
    model="llama3.2",           # Ollama model
    script_length=1000,         # Target words
    output_dir="podcasts",      # Output folder
)

agent = BlogToPodcastAgent(config)

# Convert blog to podcast
result = agent.convert("https://example.com/interesting-article")

print(f"Audio: {result['audio_path']}")
print(f"Script: {result['script_path']}")

Use Individual Components

from blog2podcasts import BlogScraper, ContentSummarizer, AudioGenerator
import asyncio

# Just scrape a blog
scraper = BlogScraper()
content = scraper.scrape("https://example.com/blog")
print(content.title, content.text)

# Just create a podcast script
summarizer = ContentSummarizer(model="llama3.2")
script = summarizer.generate_podcast_script(content.text, content.title)

# Just generate audio
generator = AudioGenerator(voice="en-US-GuyNeural")
asyncio.run(generator.generate_audio(script, "output.mp3"))

Available Voices

Recommended Podcast Voices

Voice	ID	Style
🇺🇸 Guy (Male)	en-US-GuyNeural	Professional, clear
🇺🇸 Jenny (Female)	en-US-JennyNeural	Friendly, warm
🇬🇧 Ryan (Male)	en-GB-RyanNeural	British, authoritative
🇬🇧 Sonia (Female)	en-GB-SoniaNeural	British, professional
🇦🇺 William (Male)	en-AU-WilliamNeural	Australian, casual
🇦🇺 Natasha (Female)	en-AU-NatashaNeural	Australian, friendly

Run blog2podcasts --list-voices to see all available voices.

Tech Stack

Component	Tool	Why
Scraping	Trafilatura	Best-in-class article extraction
LLM	Ollama	Free, local, private LLM inference
TTS	Edge-TTS	High-quality, free Microsoft voices
Voice Cloning	Coqui TTS	Open-source XTTS-v2 voice cloning
YouTube Download	yt-dlp	Extract audio from YouTube videos

Project Structure

blog2podcasts/
├── pyproject.toml        # Package configuration
├── LICENSE              # MIT License
├── README.md            # This file
├── CHANGELOG.md         # Version history
├── blog2podcasts/
│   ├── __init__.py      # Package exports
│   ├── cli.py           # Command-line interface
│   ├── scraper.py       # Blog content extraction
│   ├── summarizer.py    # LLM-based script generation
│   ├── audio_generator.py # Text-to-speech (Edge TTS)
│   └── voice_cloner.py  # YouTube voice extraction & cloning
├── voices/              # Saved voice profiles
└── output/              # Generated podcasts

How It Works

┌─────────────┐    ┌─────────────┐    ┌─────────────┐    ┌─────────────┐
│   Blog URL  │ -> │   Scraper   │ -> │ Summarizer  │ -> │  Edge TTS   │
│             │    │ (trafilatura)│    │  (Ollama)   │    │             │
└─────────────┘    └─────────────┘    └─────────────┘    └─────────────┘
                          │                  │                  │
                          v                  v                  v
                    Blog Content      Podcast Script      Audio (.mp3)

Troubleshooting

"Ollama not available"

# Start Ollama service
ollama serve

# Check if running
curl http://localhost:11434/api/tags

"Model not found"

# Pull the model
ollama pull llama3.2

# List available models
ollama list

"Content extraction failed"

• Some sites block scraping - try a different blog
• Check if the URL is accessible
• The fallback scraper will try BeautifulSoup

License

MIT License - Use freely for personal and commercial projects.

Contributing

Pull requests welcome! See CONTRIBUTING.md for guidelines.

Links

• PyPI Package
• GitHub Repository
• Issue Tracker