FreshScan

Inspiration

FreshScan was inspired by a common everyday frustration—despite having a refrigerator full of food, we often don’t know what’s actually inside it. This leads to food spoilage, unnecessary grocery purchases, and inefficient meal planning. We wanted to bridge the gap between physical kitchen storage and digital intelligence by giving refrigerators the ability to see, understand, and act on their contents.

What it does

FreshScan is an AI-powered smart kitchen system that scans refrigerator contents and converts them into real-time digital inventory. It automatically detects food items, recommends recipes based on available ingredients, predicts restocking needs, and generates smart grocery lists. With automation integration, it can even place grocery orders through quick-commerce platforms, reducing manual effort and food waste.

How we built it

We built FreshScan using a combination of computer vision, generative AI, and intelligent agents. A Raspberry Pi camera captures fridge images, which are processed using AI vision for food detection. The detected items are stored in a cloud database (MongoDB Atlas). An AI agent analyzes inventory, user preferences, and consumption patterns to generate recipes and grocery recommendations. Python handles the core logic, while automation tools manage orchestration and ordering workflows.

Challenges we ran into

One of the biggest challenges was achieving reliable food detection under varying lighting conditions and item occlusions inside a fridge. Maintaining an accurate and duplicate-free real-time inventory was also difficult. Integrating grocery ordering without official APIs required careful browser automation and robust error handling.

Accomplishments that we're proud of

We successfully built an end-to-end intelligent system that functions as an autonomous agent rather than a simple chatbot. FreshScan seamlessly integrates hardware, AI perception, cloud databases, and automation into a unified workflow that reduces user effort and food waste.

What we learned

This project helped us understand how to apply computer vision in real-world environments, design intelligent agent workflows, and build scalable AI-driven architectures. We also learned the importance of designing AI systems that prioritize usability and automation over manual interaction.

What's next for FreshScan

We plan to enhance FreshScan with expiry-date prediction, advanced consumption analytics, and deeper personalization based on dietary preferences. Future goals include direct integration with grocery platforms and deploying FreshScan as a scalable SaaS solution for smart homes.