audio-feeder is a Flask-based web-app that hosts your audiobooks (or other audio content) as RSS feeds compatible with podcatchers.
The main interface lists all your audiobooks once the database is updated (either by using audio-feeder update or by visiting the /update url. The default interface should also work well on mobile:
For each entry, audio-feeder can also generate "rendered" feeds, depending on what metadata is available:
- Single file: This feed consists of a single file; if the original directory contains multiple files, they will be merged together using
ffmpeg. If chapter information is not available in the original files, each file will be considered a separate "chapter". - Chapters: This feed has one file per chapter; it is only available if explicit chapter metadata is available in the file (e.g. in an
m4bfile, or using Overdrive MediaMarkers). - Segmented: This is a feed that splits up and recombines files in such a way as to create files that are ~1 hour long (preferring longer files to shorter files). Files will only be split up along chapter boundaries, but they can be combined from files without chapter information. This mode is available if the original entry has more than one file, or if it has chapter information (or both).
Each feed has an associated QR code, so that you can easily browse the list from a computer, but scan the individual feeds from your phone. In the default front-end, QR codes are displayed via modal pop-ups:
The user can customize things like the sort order and the number of entries per page:
Probably the easiest way to deploy this is via docker. For your convenience, I have created a repository with a docker-compose configuration. It uses nginx to serve static media files, and gunicorn to deploy the application.
If you want to test it out locally, you can use tox -e start_server to run a test server. This will create a temporary server directory at /tmp/audio_feeder_server, and you can play around with, and you can mount your audiobook directory at /tmp/audio_feeder_server/static/media to have it work on your own audiobooks (bind mount, symlink or hard link should work just fine).
In a virtualenv with audio-feeder installed, you can run audio-feeder --help to see the various command line tools bundled with the application.
While the state of this is getting much closer to something production-ready, I am emphatically not a front-end developer, and I don't have extremely high confidence in the security of this project. Use at your own risk (and if you do come up with some security and/or usability improvements, please do send a PR).
This is only tested on Linux, but it may also work on other platforms.
In addition to the python dependencies specified in pyproject.toml, this project also requires installing ffmpeg with at least the aac codec, and the tests also require the libmp3lame codec.
All images and documentation contained herein are licensed under CC-0.
The code is released under the Apache 2.0 license.
Pull requests and issues are more than welcome. Please be aware that your contributions will be released under the licenses stated above. If you are not comfortable with that, please do not make a pull request.