Earkind

Earkind is an AI tool designed to generate podcasts using advanced neural expressive text-to-speech technology and programmatic audio editing techniques. This innovative tool allows users to effortlessly create complete podcast episodes and descriptions based on a selection of news articles and research papers. By simply providing a .txt file containing plain text and a list of recent Arxiv papers as URLs, Earkind is able to crawl the Arxiv database, extract relevant information from the raw PDF text, and utilize the powerful ChatGPT API to generate engaging content.

One of the standout features of Earkind is its ability to define unique characters for the podcast. Hosted by Giovanni Pete Tizzano, an enthusiastic and earnest tech enthusiast, along with the sarcastic analyst Robert and the witty research expert Belinda, the podcast takes the form of an organized conversation between these characters. Scripts are created for each section, and the program parses them accordingly, ensuring a seamless and engaging listening experience.

After the recordings are completed, Earkind offers a range of jingles, sound effects, and background music options to enhance the podcast. Leveraging the capabilities of Pydub, the tool automatically adjusts volumes, overlays sections, and combines narrations and audio elements, resulting in a polished and professional final product.

In addition to the podcast itself, Earkind also generates timestamps and descriptions for each episode, making it easy for listeners to navigate through the content. The titles and overall description are also generated by ChatGPT, ensuring consistency and coherence throughout the podcast.

Earkind is not just a tool for creating podcasts, but also a platform for personalized audio content. The team behind Earkind is actively seeking feedback and suggestions from users, and they have plans to make the code public, providing an in-depth explanation of the design thinking behind this innovative tool.