Welcome to the NZRT Wiki Podcast. Today we’re looking at Podcast Folder Structure.

So, if you’ve ever wondered where all the NZRT Wiki Podcast audio files, transcripts, animated videos, and RSS feeds actually live, this episode answers that. Everything is stored in a structured folder inside Nextcloud, which is NZRT’s cloud file system hosted at cloud.nzrtnetwork.com. If you’re working locally, that same folder syncs down to your machine under your Nextcloud2 drive, inside Shared, then Content, then Podcasts.

Let’s talk about what that Podcasts folder actually looks like when you open it up. At the top level, the first thing you’ll notice is a file called feed.xml. That’s the master RSS feed — it covers every episode across every service. Sitting alongside it are a bunch of per-service feeds, each named with a service code. So you’ve got one for the x402 Knowledge Base API, one for Dolibarr, one for Blockchain, one for Flarum, one for GitHub, one for the LLM work, one for Nextcloud, one for Obsidian, one for WordPress, and so on. Each of those feeds covers only the episodes belonging to that particular service. There’s also a HOME.md file in there, which acts as an index document for the whole folder.

Below all those top-level files, you’ll find subfolders — one for each service code. So there’s a folder for 000API, one for 000DOL, one for 000BCH, one for 000NCL, and so on. Inside each subfolder you’ll find the actual episode files for that service. As an example, the 000API folder contains an mp3 file for the x402 overview episode, a txt file which is the full transcript for that same episode, and an mp4 file which is an animated visual companion for it.

That brings us to the three file types you’ll find for each episode. First, you have the mp3 file — that’s the actual audio, generated using a text-to-speech tool with an American English voice called Guy Neural. Second, there’s a txt file containing the full narration transcript — essentially the script you’re hearing right now, in text form. Third, some episodes also have an mp4 file, which is an animated visual version, created where that’s been set up for a particular episode.

Now, how do these audio files get shared publicly so podcast clients can actually play them? Each mp3 file gets a public, read-only share link created through Nextcloud’s sharing system. That link is set up as a public link with read-only access. The direct download URL from that share is then inserted into the RSS feed as what’s called an enclosure tag — that’s the standard way podcast apps know where to fetch the audio file from.

On the feed generation side, there’s a Python script called create_podcast.py that handles updating both the master feed and the relevant per-service feed each time a new episode is produced. Those feeds use the iTunes podcast namespace, which means they’re compatible with standard podcast apps. If you need to process a whole batch of wiki pages at once, there’s also a batch version of that script called create_podcast_batch.py, which will work through all the markdown files in a given wiki subfolder in one go. Both scripts live in the 000AGT repository on your local machine.

That’s really the full picture. A clean root folder with a master feed and per-service feeds, subfolders organised by service code, three file types per episode, public share links powering the RSS enclosures, and a pair of Python scripts keeping everything up to date.

That’s it for this episode of the NZRT Wiki Podcast. Thanks for listening.