Auxify is a shared Spotify play queue that allows your friends and family to queue songs of their choosing onto your queue!
Origin story
A little while ago, I was listening to music on Spotify through a bluetooth speaker with some friends. We were passing around my phone whenever someone wanted to enqueue a song of their choice, but it pained me that there wasn’t a better way, a way for everyone to add songs without needing to physically take my smartphone. That’s when I started working on Auxify. Just sharing a room ID and a room code allows people to joom my room and enqueue songs of their choice onto the play queue whenever they want. Since then, Spotify released the group session feature which fills the same niche, but I finished and launched the site regardless. Auxify is available here.
What’s in a name?
The name is a portmanteau of “aux cord” and “Spotify” - the idea being users virtually pass around an aux cord, connected to one Spotify account.
The tech
You can view the frontend source here and the backend source here
The code is split into separate repositories for the backend and frontend.
The backend
Auxify’s backend is a Python REST API, written to target Python 3.8. It makes extensive use of Python’s asyncio facility for asynchronous programming. I used the aiohttp web framework, a stable, asyncio-first web framework with a clean and easily useable interface. The deployed site uses gunicorn as the WSGI server; both the backend and frontend have NGINX as a reverse proxy in front. Most of the utilities used in the backend are provided directly by aiohttp.
For the persistence layer, I settled on sqlite3. Initially, I was working with MySQL, but I wanted to avoid unnecessary additional complexity on a website I don’t expect to achieve very many concurrent users. Websites with far greater performance requirements successfully use sqlite for persistence, and avoiding the additional complexity and operational overhead and/or cost of a MySQL server seemed worth the eventual, very distant scale barrier, especially given the app’s very low reliance on persistence. I use the aiosqlite library to provide an asyncio layer on top of Python’s built-in sqlite3 library.
The frontend
I was a little more experimental with Auxify’s frontend. I had worked with React in a professional capacity at takealot.com, but I had never used anything other than the Semantic UI framework, and hadn’t touched redux. I decided to use this as an opportunity to venture further into the JS ecosystem. I settled on Grommet as a UI framework - easy-to-use, low overhead, support for theming, flexibility and good-looking UI components out of the box, it satisfied all the requirements I was looking for in a UI framework. Axios took a lot of the overhead away from managing network requests. Redux made state management and testing components a breeze.
I still intend to investigate adding TypeScript to the codebase, but given the size of the project, it might be more hassle than it’s worth.
Future directions
I would like to add an integrated player using Spotify’s web playback SDK, specifically to give user’s a view into the current state of the queue. The idea would be to put this onto a shared screen like a TV for everyone to see.
An interesting, but legally tricky, idea is to monetize the app by using it as a sort of “jukebox” - a bar or similar establishment can allow users to use credits to enqueue songs from a predetermined list of the establishment’s choosing.