WARRGHH
Projects
BuildingWeb appIndependent

Nose Codes

Fragrance discovery — decode what you love and find your next scent.

A fragrance-discovery app that learns what you love from the notes and accords you react to, then recommends your next scent — an in-app ML recommendation as the single 'wow'. Sits on a canonical, AI-enriched ingredient database.

The problem

Fragrance discovery is overwhelming and gatekept by jargon. Nose Codes turns 'what do I actually like?' into a readable code and a recommendation you trust.

Key features

  • Decode your taste from notes & accords
  • In-app ML scent recommendation
  • Lab-notebook visual language
  • Save and compare scents

Built with — how & why

  • TypeScript
    TypeScript
  • FastAPI
    FastAPI
  • PostgreSQL
    PostgreSQL
  • Python
    Python
  • OpenAI
    OpenAI