Combina

An AI stylist that recommends outfits and finds nearby stores based on your body type, skin tone, and preferences.

Comment

Inspiration

Shopping for clothes is frustrating for many people. Even when styles look good on mannequins or models, it’s hard to know whether the same clothes will suit your body type, skin tone, or personal style. In physical stores, people often spend hours browsing without confidence, and online shopping lacks real personalization.

We wanted to solve this problem by building an AI system that thinks like a personal stylist — one that understands the user first, then guides them toward outfits and nearby products that actually suit them.

That idea became Combina.

What it does

Combina is an AI-powered personal outfit and store finder.

It starts with a short, chat-based onboarding where the AI learns about the user:

  • Name
  • Body measurements (height, weight)
  • Skin tone
  • Fashion preferences

Using this profile, Combina:

  • Recommends outfits that suit the user’s body and skin tone
  • Explains why an outfit works
  • Allows users to scan clothing items or store QR codes
  • Evaluates whether an item suits the user
  • Shows nearby stores where similar items are available (simulated for demo)
  • Includes an AI chat assistant for styling questions
  • Provides a demo “try-on” experience for visual preview

The goal is to reduce decision fatigue and help users confidently choose what to wear and where to buy it.

How we built it

Combina was built as a multi-step AI system, not just a simple chatbot.

Key components include:

  • Gemini as the reasoning engine for personalization and explanations
  • Chat-based onboarding to create a structured user profile
  • Image analysis for detecting clothing color and type
  • AI scoring logic to evaluate outfit suitability
  • Simulated store discovery using location-based reasoning
  • Modern mobile-first UI designed for quick interactions

For the hackathon prototype, store inventory and try-on features are simulated to demonstrate the concept clearly without claiming real-time partnerships.

Challenges we ran into

  • Designing AI logic that feels helpful, not overwhelming
  • Ensuring the app flow never breaks even when AI responses fail
  • Handling slow previews and session limits during development
  • Balancing explainability with fast responses
  • Making the demo reliable without a full backend

Each challenge pushed us to simplify, harden the logic, and focus on clarity over complexity.

What we learned

  • AI is most useful when it guides decisions, not replaces them
  • Explainability builds user trust
  • Simple, stable demos score higher than over-engineered systems
  • Personalization is more about reasoning than raw data

What’s next for Combina

Future versions of Combina could include:

  • Real store inventory integrations
  • Full virtual try-on using advanced vision models
  • Saved outfit history and wardrobe analytics
  • Social sharing and recommendations
  • Expansion into different cultural and regional fashion needs

Combina is designed as a foundation for a smarter, more confident shopping experience.

Built With

Share this project:

Updates