What is this?

Encouraged by friends in the Tokyo Tech community, this podcast uses AI to generate natural English summaries of NHK news stories, which are then converted into human-sounding speech. The result is short, easy-to-digest English news updates that are directly from a Japanese source.

Simiarly, I also made a site and podcast providing tech news from Japan (TotemoTech).

Technology used

• Website: GitHub pages and the Jekyll static site generator
• Theme: LinkRot podcast theme
• Audio file hosting: Internet Archive
• News source: NHK RSS feeds
• Currency rates source: Frankfurter
• English-language summaries: OpenAI API (GPT)
• Text-to-speech: ElevenLabs API – see my ElevenLabs API Python tutorial
• Programming: Good ol’ Python

The code is open source (MIT license) and hosted on GitHub. I also have a video course where I explain how to generate the audio files: ChatGPT-to-Speech: Build an AI-Powered Podcast With Python