Auxie

Inspiration

Music changes how a moment feels, but our apps have no idea what moment we're in. They'll play the same playlist whether you're cramming before a midterm, driving home at midnight, or worn out after a day of back-to-back meetings. When you skip a song, the app just plays the next one and learns nothing from it. We wanted a DJ that pays attention to your day and gets to know your taste the way a friend would.

What it does

Auxie watches your context without you doing anything: the time of day, where you are, how fast you're moving, what's coming up on your calendar, how busy your day has been, and the weather. From all that, it works out the vibe you need and queues the right song on your real Spotify. On a late-night freeway drive it picks up the energy. With an exam coming up it switches to focus music. After a draining day it brings things down to something warm. You never tell it your mood, because it figures that out on its own.

Every skip teaches it something. Auxie keeps track of which songs you skip and which ones you listen to all the way through, in each kind of moment, and uses those patterns to retrain its own model. The longer you listen, the closer it gets to your actual taste. It runs on a small open model we fine-tuned ourselves, so your location and calendar stay private.

What's next for Auxie

More signals: heart rate and workouts from your watch, screen time, and sleep, so it can tell a tired morning from a hyped one.
Fully on-device: running the model locally so none of your data ever leaves your phone.
A voice: a short DJ intro between songs that reacts to your day.
Everyone's own DJ: opening it up to more users, with each person's model learning only from their own listening.