Look closely.
Screenshot the interface until the repeats start showing up.
I used AI to structure a Figma study of variables, tokens, text styles, and components.
Independent project. Not official Spotify work or Spotify documentation.
Take an interface I use every day, work out which decisions in it repeat, and rebuild those as something reusable. I reviewed public patterns, used AI to structure the naming, and assembled the Figma file myself.
Screenshot the interface until the repeats start showing up.
Separate primitives, semantics, variables, and components.
Check that changing one token moves what it should, and nothing it shouldn’t.
Spotify uses Inter in the product. This study looks at how its hierarchy, scale and rhythm support the interface.
Every fill, border, radius and label color points at a variable, so a theme change moves all of them at once.

Play, pause, skip, progress — the pieces that have to look identical everywhere they appear.

Nothing here is used directly. Components reach the semantic layer, and the semantic layer reaches this.

Spacing, radii and sizes as numbered steps, so a gap is chosen from the scale instead of typed into the inspector.

Families, weights and sizes as a pool. The named text styles below draw from it.

Surface, text, border, accent. Renaming a primitive never reaches this far, which is the whole point of the split.

Container padding and stack gaps mapped onto the number primitives, so layout is described rather than measured.

If a size isn’t one of these seven, it isn’t in the system. That constraint is what keeps the interface readable at a glance.







Song and artist cards compared at rest, on hover, and once selected.






AI helped organize patterns, tokens, names, and consistency checks. I made the design decisions and assembled the Figma studies.
Independent Spotify-inspired study. Not official Spotify work or documentation.