Skip to content

Latest commit

 

History

History
102 lines (72 loc) · 3.09 KB

README.md

File metadata and controls

102 lines (72 loc) · 3.09 KB
Raw

AI Podcast Generator

This script generates an AI-powered podcast based on a given text. It transforms a conversation between two AI personalities into a fully synthesized audio file. You can customize the speakers, language, and voices for your podcast. Inspired by the audio overview option of Google's NotebookLM experiment, and inspired by AnthusAI/Podcastic to make a script of it.

Animated gif of the script in action

Requirements

Before running the script, make sure you have the following installed/configured:

Installation

macOS/Linux

  1. Clone the repository:

    git clone https://github.com/timonvanhasselt/AI-podcast-generator.git
cd ai-podcast-generator

  2. Create a virtual environment:

python3 -m venv venv
source venv/bin/activate
  1. Install the required Python packages: pip install -r requirements.txt

Windows

  1. Clone the repository:
cd ai-podcast-generator
  1. Create a virtual environment:
python -m venv venv
venv\Scripts\activate
  1. Install the required Python packages:

pip install -r requirements.txt

Usage

Update the Hugging Face credentials in the script to your own email and password:

EMAIL = "[email protected]"
PASSWD = "your-password"

Prepare your input: The script reads text input from a file named content.txt. Create a content.txt file in the project directory and input the text you want to be transformed into dialogue.

Run the script: Once everything is set up, you can run the script as follows:

On macOS/Linux: python3 podcast_script.py

On Windows: python podcast_script.py

Output

The script will generate an SSML conversation between the speakers, convert it into speech, and combine the segments into a final MP3 file named final_output.mp3.

Customization

Change Speakers and Language You can customize the speakers and their voices by modifying the following variables in the script:

speaker1 = "Ava"  # Change to your preferred speaker name
speaker2 = "Andrew"  # Change to your preferred speaker name
lang = "English"  # Change to your preferred language

You can also map different voices using EdgeTTS by updating the voice_map dictionary:

voice_map = {
    "Ava": "en-US-AvaMultilingualNeural",  # Change to your desired voice from EdgeTTS
    "Andrew": "en-US-AndrewMultilingualNeural"
}

Notes

Make sure you are logged into Hugging Face and have the proper credentials for API access. If you encounter any issues with model overload, wait and try again later.