Beta testing
Moodify
A mood-first music app I started after noticing that a song can feel right before I know how to describe my mood. I’m testing how recommendations can learn from listening without pretending feelings are simple.
Demo
A short walkthrough of Moodify, from choosing a mood to personalized playback.
This shows the current Moodify flow from mood selection through playback and feedback.
- What I noticed
- When I listen to music, I do not always start with a clear mood name. Sometimes I just know that a song fits, or that it does not. I wanted to understand whether an app could learn from that kind of signal without pretending feelings are simple.
- What I tried
- I built an iOS app where the user starts from how they feel, then the app recommends music and adjusts over time based on listening behavior and feedback.
- What I worked on
- I designed the app idea, built the iOS interface, worked on recommendation logic, tested playback behavior, and kept improving parts that felt confusing or unreliable.
- What I used
- Swift, SwiftUI, Apple Music integration, recommendation ranking, per-mood learning, playback queue logic, and local diagnostics.
- What I learned
- Mood is harder than a button. A person can choose one feeling, skip a song, replay another, and still not be easy to understand. I learned that a recommendation system needs to be careful about what it thinks it knows.
- Where it is now
- Moodify is available as a TestFlight beta.