TryRoom

• 

• 

• 

• 

• 

• 

• 

• 

  icon

Inspiration

What it does

How we built it

Challenges we ran into

Accomplishments that we're proud of

Inspiration

We've all had that moment of staring into a full closet and feeling like there's "nothing to wear." At the same time, we keep buying new clothes that end up sitting unworn because they don't match anything we already own. We wanted to fix both problems at once — not by telling people to buy more, but by helping them actually use what they already have, and shop smarter when they do buy. That's the idea behind TryRoom: a personal AI stylist that knows your entire wardrobe and can dress you for any occasion, instantly.

What it does

TryRoom is a personal AI stylist that lives in your pocket. Users photograph their clothes to build a virtual wardrobe, and AI analyzes each item's color, style, pattern, and season. From there, TryRoom:

• Matches items already owned into AI-curated outfit combinations
• Recommends outfits based on occasion (job interview, date, wedding, casual Friday, etc.)
• Lets users photograph items while shopping to preview how they'd pair with their existing wardrobe before buying
• Visualizes complete outfits on AI models or the user's own photo

How we built it

TryRoom is a React Native mobile app, built with the help of Claude Code. The core of the product is an AI pipeline that handles style analysis (breaking down each photographed item into attributes like color, pattern, category, and season) and style generation (combining those items into complete, coherent outfits).

Challenges we ran into

• Matching logic across styles and seasons — building a system that understands not just "does this color go with that color" but whether a full outfit reads as stylistically coherent and season-appropriate was far more nuanced than simple rule-based matching.
• Multi-item attachment for complete outfit generation — getting the LLM to reason over a full combination of items at once (shoes, clothing, watch, accessories) rather than pairing items one at a two at a time, so the output is a genuinely complete, styled look rather than a partial suggestion.

Accomplishments that we're proud of

We're proud that TryRoom doesn't just match colors — it reasons about full outfits the way a real stylist would, pulling together multiple wardrobe items into a single cohesive look. Getting an agentic pipeline to reliably handle that level of multi-item reasoning, on a mobile app built in hackathon time, felt like a real win.

What we learned

We learned how powerful agentic pipelines can be for creative, subjective tasks like styling. Instead of a single AI call generating an outfit in one shot, we built a pipeline where agents refine the style across multiple rounds — critiquing and improving the combination before presenting it. That iterative, agent-driven approach produced noticeably better and more coherent outfit quality than a single-pass generation ever could.

What's next for TryRoom

• Let users share clothes they no longer wear, so others can browse and claim them
• Explore rental and second-hand resale options directly in the app
• Let stores list their inventory so users can combine store items with their own wardrobe before buying ## What we learned

What's next for TryRoom

Built With

• claude
• reactnative