HomeWorkAboutContact
← Back to work
BuiltPassion

Moodify

What do you listen to when you know how you feel, but not what you want to hear? Moodify turns moods into playlists. Pick up to three emotions and get music that matches your state of mind.

3 min read
Live Preview ↗
Stack
Next.js · Spotify API · TypeScript
Year
2026
Role
Designer & Developer
Status
Live
The Spark

Spotify is built for people who know what they want. Who follow artists, remember album names, know what genre fits a Tuesday evening. I don't always have that. I have feelings. Most days I know I need something calm, or something sitting between nostalgic and hopeful, but I couldn't tell you which song gets me there. I'd open Spotify, stare at it, feel more frustrated than before, and give up.

I also thought about people who aren't music obsessives. Older users who don't follow new releases. People whose Liked Songs playlist is full of tracks they saved without ever checking the artist's name. Anyone who wants music to match a feeling but doesn't have the vocabulary to search for it.

In Context
Decisions
02
Sort by match score

Spotify returns recommendations in its own order. Resorting by the calculated score means the tracks most aligned with your mood come first, not whatever Spotify's algorithm surfaces at the top.

03
Three track states

Once a track plays, it dims to greyscale and shows a checkmark. A small thing, but it makes a long playlist feel navigable. You always know where you are. The greyscale treatment also gives the playlist a sense of time passing. You can see your session taking shape.

04
Play locked until you commit

The Play button stays disabled until at least one mood is chosen. When three moods are selected, unselected tiles dim. Both are small friction points, intentional ones. The interface asks you to decide before it plays.

What Worked
01

The core experience: pick a mood, hear matched music within seconds

02

Share generates a social card: a PNG snapshot of the mood moment

03

Music discovery: surfaces artists you wouldn't find through normal browsing

04

Multi-mood combinations return something genuinely different from either mood alone

In Progress
01

Save to Spotify has an OAuth permissions scope issue currently being resolved

02

Multi-mood blending for emotionally distant combinations needs a smarter algorithm

What's Next
01

Sequence playlists to gradually shift the listener toward where they want to be, not just mirror the current mood

02

Smarter multi-mood blending that weights emotionally distant combinations differently, instead of averaging opposing targets

03

Surface patterns in how mood and listening habits shift over time

04

A living share card someone else can actually play, not just save as a PNG

Got a project in mind?

I'm available for new projects.

Start a conversation →
← Previous
Next →