Set up your own automated AI-driven news-monitoring podcast
Micke Kring
·
·
2 min read
Keeping up with what’s happening. Not always easy to find the time for it. And how am I supposed to know if that 20-page report is for me? And how am I supposed to check all those sites every other day? That was the starting problem for me with this little prototype.
Imagine if an application could keep track of all the news that appear on my favorite sites, pick out the interesting ones based on keywords I’ve specified, have an AI read those articles and summarize them into one minute of spoken text, have another AI narrate those, and finally create an entire podcast from it. Well, that’s what ARAIA is. An Automated, Robotized and AI-Assisted podcast engine. :)
Workflow
ARAIA, which you can also tinker with now that I’ve posted the code on GitHub, does the following:
- Reads the news feeds you’ve chosen and looks for the keywords you’ve specified in the various articles.
- If there’s a match in step 1, the full article is sent to GPT-4 for “reading and summarizing” into a format that can be read aloud in one minute. Everything is stored in a database.
- When there are enough articles, the summaries are sent to Microsoft’s TTS, which converts the text to speech.
- Podcast episodes are created by mixing intro, outro, background music, and the articles into an episode.
- The website that ARAIA has created is updated with the latest episode.
- The podcast feed, the XML file, is updated.
- Services like Apple, Google and Spotify Podcasts pick up your episode.
In the video below, which is the last in a series of videos where I developed this, I show how everything works.
Want to run it yourself?
On GitHub I’ve uploaded all the code and a simple guide on how to get started. github.com/mickekring/ARAIA-Podcast-Engine
Want to see an example - Svarta Tavlan
If you want to see a “real” podcast using ARAIA you can go to svartatavlan.mickekring.se. It’s a podcast sharing news about the digitalization of schools.
Ideas for use cases
- Your very own tailored podcast
- Your company’s news to your employees
- Your own website’s podcast
Questions?
As usual, I try to answer any questions you might have. Comment on the post, or get in touch via social media.