Inspiration
50% of Americans have prediabetes or diabetes. 75% of adults have at least one chronic condition. 70% of an American child's diet is ultra-processed.
120 billion pounds of food are wasted per year in America. This accounts for a loss of 2,400 million GJ (664 billion kWh) of energy – enough to power more than 50 million U.S. homes for a year; and 170 million metric tons of CO2e GHG emissions (excluding landfill emissions).
What it does
Sift aims to solve this problem at both an individual and commercial level.
For individuals, we recommend healthier, climate-and allergen-friendly alternatives to many foods and offer healthy recipes using your current pantry when food is about to spoil, ensuring no food goes to waste.
On an industrial level, Sift can be used in commercial kitchens to make the food served healthier, more carbon-conscious, and less wasteful.
How we built it
Sift is built using Golang, Next.js, Neon, and the Gemini API. Gemini is used for auto-categorization, shelf-life prediction, and recipe suggestions, and our food data is sourced from the OpenFoodFacts API.
Challenges we ran into
Using the OpenFoodFacts API was somewhat difficult; their dataset is massive and noisy, making it hard to figure out what is relevant to us and what isn't. We ended up filtering the 4.5 million data points down to ~70,000, ensuring only relevant, US-based products are recommended to the consumer.
Accomplishments that we're proud of
We believe Sift's notifications and recipe generation for items expiring soon can truly have a massive impact in cutting down on food waste.
What's next for Sift
We would love to add a local supermarket/farmer's market API to better inform individuals and commercial customers about the alternative products available in their area.
Log in or sign up for Devpost to join the conversation.