YARN - Sven Meacham

Inspiration

I was inspired by apps like Yuka, which scan food products and break down what’s inside them with a clear score. I wanted to bring that same idea to clothing, specifically from a sustainability angle, so people can better understand what their clothes are made of and make more informed purchasing decisions.

What it does

YARN allows users to scan or upload a photo of a clothing care tag, extract key information like brand, material composition, and country of manufacture, and generate a sustainability-style score. The score is based on materials (e.g., cotton vs. polyester), brand practices (ethics, sustainability, transparency), and country of manufacture as a proxy for production conditions, and the app presents a simple, easy-to-understand breakdown explaining why the item received its score.

How we built it

YARN was built using Cursor for rapid development, OpenAI for parsing and structuring OCR text, and Supabase for storing brand, material, and scoring data. The frontend was developed with React Native using Expo. The app takes an image of a clothing tag, extracts text via OCR, structures that data, and then queries Supabase to generate a final score and explanation.

Challenges we ran into

The main challenge was getting reliable outputs from AI, especially parsing messy clothing tag text into structured data. Small inconsistencies in OCR or formatting could break downstream logic, so a lot of time went into prompt iteration and making the system more robust.

Accomplishments that we're proud of

We built a fully working demo with a functioning scanning system and explanations.

What we learned

We learned how to combine AI, OCR, and a database to build a simple but functional product quickly, and how important it is to keep the scope focused for a hackathon.

What's next for YARN - Sven Meacham

We want to improve the accuracy of tag parsing and expand the brand and materials database, while adding more personalized and detailed scoring.

Built With

• cursor
• openai
• supabase