Homepage
Sample room

Snapify

Inspiration

Staring at empty rooms and dorms and wondering how to make them feel like home. Buying furniture online only to discover it doesn't fit your space or match your vision. We wanted to bridge the gap between imagination and reality in interior design - so we made Snapify.

What it does

Snapify combines Snap's AR glasses with AI to let you see furniture in your actual space before buying. Tell it your style - modern, cozy, gothic, rustic, industrial, or traditional - and watch as curated furniture appears perfectly placed in your room. Walk around, see how pieces look from different angles, then purchase what you love.

How we built it

Snap AR Spectacles for spatial scanning and 3D object placement using World Query Module
Gemini Live API for intelligent voice interaction and real-time conversation in AR
Trellis 3D API for generating realistic furniture models from text descriptions
React + Vite + Tailwind CSS for the companion web interface
TypeScript/Lens Studio for AR application development and 3D object management
Shopify API to get furniture data and product information

Challenges we ran into

The biggest challenge was texture rendering in AR. 3D models generated from Trellis API looked perfect in preview but lost colors and appeared transparent when imported into Lens Studio. We had to simplify texture complexity, reduce resolution, and recreate materials manually to make them AR-compatible while maintaining visual distinctiveness.

Accomplishments that we're proud of

Successfully connected a mobile app with AR glasses for seamless cross-platform interaction
Built an intelligent caching system that pre-generates 3D models for instant placement
Created a natural conversation flow where users simply describe their style preference
Integrated multiple complex APIs (Gemini, Trellis, Snap AR) into 1 cohesive experience

What we learned

AR development requires different optimization strategies than traditional 3D applications. Texture complexity that works perfectly in other contexts can break AR experiences. We also learned that effective AI integration isn't just about powerful models - it's about designing prompts and data structures that bridge human preferences with technical constraints.