The problem

You have a directory of audio files that you'd like to turn into a self-hosted podcast RSS feed.

There are various reasons you might want to do this. My use-case is I have 150 audiobooks that I've purchased and I don't want to use Audible or Libro.fm (for example) to listen to them so instead I've downloaded the MP3 files. It's too much to store on my device all at once, and I want to be able to cast them to my TV/speakers with Chromecast. I also need this to be password protected to avoid stealing my library.

This solution

This is a node server which will serve an RSS feed of all the audio files in a directory of your choosing. It uses express to handle GET requests to /audiobook/feed.xml, /audiobook/:bookId/image, and audiobook/:bookId/audio.mp3 (the only one you need to ever access directly is the feed).

To solve my specific problem, I store all my audiobooks on the Synology DiskStation (NAS) I have and run this server directly on that NAS (scaling is not an issue because my family are the only ones that use it). From there, I use a podcast app (the best I've found is PocketCasts) to consume the feed and access/download all my audiobooks. Most podcast apps support reading the chapter ID3 tags embedded in the audiobook files so you even get chapter support. It's pretty seamless!

Table of Contents

• Installation
• Usage
  • Project setup
• Other exports
• Other info
  • query string
  • cache
  • Editing book metadata
• Other Solutions
• About @kentcdodds/ scoped and kcd- prefixed packages
• Issues
  • 🐛 Bugs
  • 💡 Feature Requests
• Contributors ✨
• LICENSE

Installation

This module is distributed via npm which is bundled with node and should be installed as one of your project's dependencies:

npm install --save @kentcdodds/podcastify-dir

Usage

const path = require('path')
const {startServer} = require('@kentcdodds/podcastify-dir')

startServer({
  // the title will appear in the podcast app identifying this feed
  title: 'Podcast Title',

  // the description will normally appear on the feed's screen in the podcast app
  description: 'Some great audiobooks',

  // This image will show up in the podcast app for this feed
  image: {
    url: 'https://www.example.com/some-image.png',
    link: 'https://www.example.com',
    height: 500,
    width: 500,

    // I'm not 100% certain what these are for...
    title: 'Some title for the image',
    description: 'Some description for the image',
  },

  // the port you want to bind to (if not specified, it chooses a random port)
  port: process.env.PORT,

  // the directory of audio files
  directory: path.join(__dirname, '..', 'audiobooks'),

  // the username and password that will allow you to access the feed
  users: {bob: 'the_builder'},

  // this allows you to pass your own express app so you can configure it
  // however you like. It defaults to creating one itself
  app: express(),

  // this allows you to specify where you want the routes to be mounted to
  // by default it's /audiobook, but you could change it to "/" or "/podcast"
  // if you'd like.
  mountpath: '/audiobook',

  // a little inversion of control here to allow you to modify the JS object
  // that's converted to XML. We're using the `xml-js` npm module so you'll
  // want to make sure your modifications will work with that package's
  // `convert.js2xml` method
  modifyXmlJs(xmlJs) {
    xmlJs.rss.channel['itunes:author'] = 'Your name'
    xmlJs.rss.channel['itunes:summary'] = 'Some other stuff'
    return xmlJs
  },
})

startServer returns a promise with the started server in case that's useful. It also ensures that the server is shut down properly if the process exits.

The server also supports rate limiting to help avoid people brute-forcing the username/password.

Project setup

In my project, I only need a few things to get this running: package.json, index.js, and forever.config.json

package.json:

This lists the dependencies and the scripts for the project.

{
  "private": true,
  "name": "doddsfam-audiobooks",
  "version": "1.0.0",
  "main": "index.js",
  "scripts": {
    "stop": "FOREVER_ROOT=./.forever forever stopall",
    "start": "FOREVER_ROOT=./.forever forever start ./forever.config.json"
  },
  "license": "UNLICENSED",
  "dependencies": {
    "@kentcdodds/podcastify-dir": "^1.4.2",
    "forever": "^3.0.0"
  }
}

forever.config.json:

forever is a module that will ensure that if the server stops for any reason, it is automatically restarted. That way you don't have to log into your server to restart it if it crashed. Here's how I configure it:

{
  "append": true,
  "script": "index.js",
  "sourceDir": ".",
  "logFile": "./forever.log",
  "outFile": "./out.log",
  "errFile": "./error.log"
}

index.js:

const {startServer} = require('@kentcdodds/podcastify-dir')

startServer({
  title: 'Dodds Family Audiobooks',
  description: 'The audiobooks of the Dodds family',
  image: {
    url: 'https://www.dropbox.com/s/some-id/some-filename.jpg?raw=1',
    title: 'Dodds Family Audiobooks',
    link: 'https://kentcdodds.com',
    height: 500,
    width: 500,
  },
  port: 8879,
  directory: '/volume1/audiobooks/files',
  users: {bob: 'the_builder'},
  modifyXmlJs(xmlJs) {
    xmlJs.rss.channel['itunes:author'] = 'Kent C. Dodds'
    xmlJs.rss.channel['itunes:summary'] = 'Dodds Family Audiobooks'
    return xmlJs
  },
})

Because my NAS is accessible via the world-wide-web, I can use this URL in my podcast app:

http://bob:[email protected]:8879/audiobook/feed.xml

That's:

http://[username]:[password]@[domain_or_global_static_ip_address]:[port]/audiobook/feed.xml

I paste that into PocketCasts's submit page (specifying "private" so it's not indexed) to get a pocketcast URL for the podcast and then load that up in my pocketcasts account. I've had some success with other apps, but I've had some trouble with all of them. I recommend experimenting a bit.

Other exports

getPodcastMiddleware and getPodcastRoutes are also exported if you'd like to use those more directly. I don't plan to document those, but feel free to explore the source code if you need something more advanced.

Other info

query string

The /audiobook/feed.xml endpoint allows you to filter and sort the audiobooks via the query string. You can also specify a custom image. This allows you to set up custom feeds for different categories of books. Here's a full example (put on multiple lines to simplify reading it):

http://bob:[email protected]:8879/audiobook/feed.xml
?filterIn=fantasy:category
&filterOut=poppins:title,joe:author
&sort=desc:pubDate,asc:duration
&title=Fantasy%20books
&image.url=https%3A%2F%2Fwww.dropbox.com%2Fs%2Fsome-id%2Fsome-name.png%3Fraw%3D1
&image.title=epic%20fantasy
&image.link=https%3A%2F%2Fkentcdodds.com
&image.height=500
&image.width=740
&image.description=epic%20fantasy

So you have title, filterIn, filterOut, sort, and image.* options.

The fitler query params are a list of comma-separated filter sets which is a pair of [regex]:[property]. The sort is [direction]:[property].

Additionally, the filterIn and filterOut options can be used multiple times. So if you want to filter in "self help" and "career" categories, then you could use multiple filterIn params. For both filterIn and filterOut, the item must match all of the query options to apply.

Alternatively, you could create individual feeds by starting multiple servers on different ports and putting the audio files in different directories.

cache

Because reading the audiobook MP3 files for metadata can take some time (added 300ms to the request when testing on my MacBook with just a few books), we cache the metadata in memory. This is a pretty significant perf savings. However, if you add a new book, or change metadata about a book, you'll want to delete the cache, so there's also a /audiobook/bust-cache endpoint you can hit with a GET request and it'll reset the cache.

Editing book metadata

I use Kid3 for editing book metadata. It works pretty well. If you're audiobooks come from a reputable source (most of my books come from Audible which I download using OpenAudible), all the metadata should be set properly already. If you need to edit things manually, here are the values you need to have set:

• title - book title
• comment - book summary
• asin - book ID
• artist - book author
• duration - the time duration of the audiobook
• narrated_by - the person (or people) who narrated the audiobook (optional)
• book_genre or genre - A colon-separated list of applicable categories: Kids 8-10:Adventure:Fantasy
• year - The release date of the audiobook: 2020-01-23
• APIC - The Cover art (Kid3 allows you to drag-and-drop an image).

Other Solutions

I'm not aware of any, if you are please make a pull request and add it here!

About @kentcdodds/ scoped and kcd- prefixed packages

If a package I maintain is scoped to my username (@kentcdodds) or prefixed with kcd-, that means I built and maintain it for myself. You're more than welcome to use it, but I'm not likely to put much work into making it work for other people's use cases (I'm not heartless, I just don't have the time). If you have a grander vision for the project, please feel free to bring it up in the comments and perhaps we can collaborate on that vision and make it more general-purpose (and remove the scope/prefix), but it's possible I'll recommend you just fork the project and publish your own version.

Issues

Looking to contribute? Look for the Good First Issue label.

🐛 Bugs

Please file an issue for bugs, missing documentation, or unexpected behavior.

See Bugs

💡 Feature Requests

Please file an issue to suggest new features. Vote on feature requests by adding a 👍. This helps maintainers prioritize what to work on.

See Feature Requests

Contributors ✨

Thanks goes to these people (emoji key):

This project follows the all-contributors specification. Contributions of any kind welcome!

LICENSE

MIT