Earkind

What is Earkind?

Earkind is an AI tool that produces podcasts about Artificial Intelligence, combining AI news, deep dives into research papers, and humorous content. The goal is to make AI podcasts that are both informative and entertaining. Shows are hosted by Giovani Pete Tizzano, Robert, and Belinda, including the ‘GPT Reviews’ show, and episodes are available on Spotify, Amazon Music, and Apple Podcasts. Content is tailored for anyone interested in AI, whether a casual enthusiast, researcher, or scholar. It is not yet clear whether the tool is free or subscription-based. Feedback and suggestions can be emailed to the creators at [email protected].

How Earkind works

Earkind automates the full podcast production pipeline:

• It creates podcast episodes and their descriptions from a list of news articles and arXiv papers.
• It crawls titles and abstracts from arXiv papers and can also extract information directly from raw PDF text.
• It helps create and fine-tune the system and user prompts for each part of a podcast.
• Content discussions are organized in a conversational style featuring different characters.
• Azure’s Text-to-Speech (TTS) is used to record the audio.
• Episodes are edited with jingles, sound effects, and background music, often using Pydub.
• It automatically generates detailed podcast descriptions, complete with timestamps and titles.

Who created Earkind?

Earkind was created by Sergi, who set out to produce AI-powered podcasts covering the latest AI news, research papers, and humor.

Feedback and future plans

The creators actively seek feedback and are gauging interest in offering personalized content. There are also plans to make the code public, providing transparency and an in-depth explanation of the design process.