← Back to Projects

Insights from Podcasts

In this project I built an interactive terminal app to download podcasts, create transcripts and generate insights from those podcasts. The insights can be emailed or distributed in any other way. This project is work in progress (last update: 14.10.2025).

My goal is to create an interactive terminal app that will monitor selected podcasts for new episodes and makes insights from those podcasts available in my preferred reading app. As I'm building this I'm aiming to produce independent modules that can be reused in other projects. I also hope someone else will find these modules useful and will build something with them. All of my work is published on my Github and can be freely used in other projects. I'd be happy if you reach out and let me know what you built with it.

Index

Podcast Insights via Interactive Terminal

Apps like Claude Code have proven that terminal apps can be easy to use and possess a minimalistic beauty. Inspired by this, I wanted to create my own interactive terminal app and my Podcast Insights project provided me with that opportunity.

Podcast Insights Library

Podcast Insights comes with a simple config file where you can add the shows that you want to monitor. Launching the app allows you to check the newest episodes of any show and start processing any episode you like. The processing screen then shows progress of the download, the transcription and ultimately the extraction of the insights. The original MP3 file, the transcript and the markdown file with the insights are stored in a data directory that can be configured.

Podcast Insights is available for free on my GitHub.

The app is still very actively developed and will see breaking changes on a regular basis. As for my roadmap, the next feature will be the option to configure additional actions after the insight extraction is done. For example, in order to read the insights in my favorite reading app, Readwise Reader, I need to trigger an action that converts the Markdown to HTML and then sends that HTML content via API to Readwise.

Module 1 - Local audio file transcription

Nvidia published a multilingual speed-to-text model called Parakeet, and it's open source. It was originally built for Nvidia's own CUDA architecture (which means you would need a Nvidia GPU to use it), but thanks to the parakeet-mlx project, it also works perfectly on Apple's M-series chips on a Mac. It actually runs so well that it allows you to quickly transcribe any audio file with a high level of accuracy. Compared to the transcription APIs offered by various AI vendors, Parakeet is free - you only pay for electricity (well...and the Mac).

I used Claude Code to create a little Python script that takes any audio file, transcribes it using parakeet-mlx and outputs a markdown file. This script can serve as the foundation for other, more complex use cases, like a local transcription web app (like my own EchoScribe) or generating summaries and insights with AI. The script is available on GitHub.

Local Audio Transcription

Using the script in Raycast Since I'm a big fan of Raycast, I also created a copy of the script that has the required configuration to serve as a script command in Raycast. This allows you to easily start a transcription from anywhere just by typing "transcribe" and providing the path to the file you want to transcribe. More detailed instructions for setting it up can be found in the README file inside the Github repo.

Podcast Transcription Raycast

Module 2 - AI Prompting via Command Line (AI CLI)

I'm using various AI models to extract insights from the transcribed podcasts. Instead of listening to those podcasts for many hours I will get a summary and a list of insights that I can read in a few minutes.

I was hoping to use local AI models for this job, but even on a modern M4 Pro chip from Apple those models struggle with large amounts of content due to limited memory. It's not that a local model like qwen3 didn't do the job at all - the quality of the generated insights from a large model like GPT-5 or Claude Opus was just much better (even when using the smaller models like GPT-5 mini).

Since my ultimate goal is to automate the entire process from podcast transcription to insight generation, I built a little AI CLI tool that allows me to process the podcast transcripts (or anything else really) on the command line. I uploaded the project on GitHub for everyone to use or extend: AI CLI on Github.

AI CLI