FoodFresh AI

• 

  Home page

Inspiration

In today’s fast-paced lifestyle, we often forget about the food we’ve bought until it expires. With global food waste contributing significantly to environmental issues and household expenses, we were inspired to create a solution that uses AI to track expiry dates, remind users before food spoils, and promote smarter consumption. Our goal is to simplify food management while encouraging sustainability.

What it does

FoodFresh AI helps users:

Upload food packaging images; OCR + LLMs extract expiry and manufacture dates.

Manually add fresh produce like fruits and vegetables with preset shelf lives.

Get smart reminders before food expires.

View a personalized dashboard showing food status sorted by expiry proximity.

Receive recipe suggestions based on available ingredients to minimize waste.

How we built it

Frontend: Streamlit for a fast, clean user interface.

Backend: Python with FastAPI handles API requests, user authentication, and database logic.

Image Recognition & OCR: Used Tesseract for OCR and Hugging Face LLMs to understand date context (e.g., differentiating between "MFG" and "EXP").

Database: SQLite for lightweight, fast data storage.

Recipe Engine: Integrated a rule-based matching system with available ingredients to recommend recipes.

Authentication: User registration and login system implemented using session state in Streamlit.

Challenges we ran into Extracting accurate dates from varied food package formats was difficult due to different fonts, layouts, and languages.

Distinguishing between manufacture and expiry dates using OCR alone often failed, so we incorporated LLMs for context understanding.

Streamlit’s session management required workarounds to maintain login state across pages.

Limited datasets for training and testing the recipe and expiry logic slowed development.

Accomplishments that we're proud of

Developed a functional AI-driven expiry tracker from scratch.

Successfully integrated OCR with LLMs to extract meaningful expiry/manufacture dates.

Built a clean, intuitive interface that non-technical users can easily navigate.

Created an end-to-end reminder system with food status visualization and recipe suggestions.

What we learned

Combining OCR with LLMs significantly improves the reliability of data extraction from images.

User experience is key—clear, minimalist design with smart categorization keeps users engaged.

Effective team collaboration and modular code design speed up testing and iteration.

Even a small solution can contribute meaningfully to sustainability and personal wellness.

What's next for FoodFresh AI

Mobile App Development for on-the-go scanning and reminders.

Barcode scanning to auto-fetch expiry dates from public food product databases.

Cloud storage and syncing across multiple devices.

User-customizable reminders and integration with smart assistants (like Alexa or Google Assistant).

Gamification to encourage sustainable habits (e.g., “Zero Waste Streaks”).

Language and region support for global accessibility.

Built With

• genai
• huggingface
• llm
• python
• streamlit