Clearlabel

A smart way to understand how clean your food is and how it affects you.

Comment

Inspiration

Most packaged foods today contain long, confusing ingredient lists filled with unfamiliar names. Even when labels exist, understanding what’s actually healthy requires time, research, and nutrition knowledge. Inspired by EU nutrition ratings and the lack of truly clean ingredient options in the market, we wanted to build something that helps users decide before struggling to decode labels.

What it does

Clearlabel lets users scan or upload a food package image to instantly analyze ingredient quality. It evaluates cleanliness (ingredient count), artificial additives, emulsifiers, sugar, salt, and nutritional balance. The app personalizes results using blood reports, medical conditions, and allergies, flags unsafe foods, recommends healthier unbiased swaps, and provides analytics on shopping habits.

How we built it

Frontend: React with live camera scanning and image upload. Backend: FastAPI with Python, integrating Google Gemini for text/vision extraction, OpenFoodFacts API, and custom scoring logic. Data storage & analytics: Snowflake for user profiles, scans, and analytics (first time using Snowflake was both exciting and challenging). Process: Initially experimented with web scraping grocery data, then transitioned to reliable public food databases for faster, more accurate ingredient retrieval. Health logic: Combines Nutri-Score, WHO daily recommendations, and traffic-light labeling for personalized scoring.

Challenges we ran into

Accurately extracting ingredient data from packaging images was challenging, especially optimizing for speed and reliability. We initially explored web scraping grocery platforms before transitioning to public databases. Integrating Snowflake for the first time was both exciting and difficult, particularly designing analytics workflows efficiently.

Accomplishments that we're proud of

We built a system that delivers meaningful health insights in seconds, often faster than reading the first two ingredients on a package. Combining global nutrition standards with personal health data into one seamless experience was a major milestone. The unbiased recommendation engine was another highlight.

What we learned

We gained hands-on experience working with real-world food data, public nutrition standards, and analytics pipelines. We also learned how critical performance, data normalization, and personalization are when building health-focused consumer applications.

What's next for Clearlabel

We plan to enhance recommendations based on the specific store a user is shopping in, including private-label products. Future plans include browser extensions for online grocery platforms, deeper personalization, and expanded coverage across global food databases.

Built With

Share this project:

Updates