FridgeAI is a cutting-edge smart kitchen assistant that makes cooking easier. Its AI-powered features recommend recipes based on your mood and the items you have on hand, and it updates a virtual fridge by scanning receipts to make sure you're always aware of what you have.

Inspiration

The goal of minimizing food waste and simplifying meal preparation gave rise to the concept of FridgeAI. We saw how difficult it may be to keep track of groceries and find recipes, so we had an idea for a tool that would make regular cooking a fun, tech-savvy activity.

What it does

  • Receipt Scanning & Virtual Fridge: Automatically update your inventory by scanning receipts.
  • AI-Driven Recipe Recommendations: Suggest recipes based on your mood and available ingredients.
  • Interactive Cooking Experience: Offers text-to-speech guidance with voice commands to navigate through cooking instructions.
  • Gamified Cooking: Engages users with a “streak” feature that rewards consistent cooking and kitchen adventures.

How we built it

We leveraged a mix of cutting-edge technologies and frameworks to bring FridgeAI to life:

OCR & Data Extraction: We integrated Google Cloud Document AI alongside OpenCV to accurately scan receipts and extract relevant data. Normalization function: Dealing with different versions of a word. Example: Milk, Mlk are the same. RAG Recipe Recommendation: Pinecone powers our vector database, acting as our AI agent to recommend food dishes based on available ingredients and user preferences. Tech Stack Structure: The front end was built with React and styled with Tailwind CSS, while FastAPI handles our backend processes. Deployment & Containerization: Our application is deployed on Render, Vercel, and containerized using Docker to ensure smooth scalability and reliability. Data Manipulation: We utilized NLP libraries like SpaCy to process and analyze data, ensuring that our recommendations and insights are accurate and personalized. This robust combination of technologies ensures that FridgeAI works seamlessly and scales to meet our users' evolving needs.

Challenges we ran into

Ideation: Thinking of an idea for the project to get us started. Accuracy in Receipt Scanning: Ensuring precise data extraction from various receipt formats. Complex Recipe Matching: Balancing user mood with the ingredient availability to generate appealing recipes. Updating Fridge Content: Update the content of the virtual fridge according to the user's scan receipt. User Engagement: Designing a gamified system that genuinely motivates users without feeling gimmicky.

Accomplishments that we're proud of

Seamless Integration: Merging receipt scanning, AI recipe recommendation, and voice commands into a single, cohesive platform. User-Centric Design: Creating an intuitive interface that caters to both tech-savvy cooks and culinary novices. Innovative Gamification: Successfully incorporating a streak feature that has driven repeated user engagement and enthusiasm.

What we learned

Our knowledge of machine learning applications in daily life and the value of customer feedback in improving tech goods has grown as a result of this research. We also learned that combining several cutting-edge technologies, such as voice recognition, NLP, and OCR, necessitates a delicate balancing act between creativity and pragmatism.

What's next for FridgeAI

In the future, we intend to increase FridgeAI's functionality by: Improving AI Personalization: Adjusting recipe suggestions to more accurately represent unique cooking methods and nutritional requirements. Connecting kitchen devices to enable more automated cooking is known as "smart appliance integration." Growing Language Support: Expanding the availability of recipes and voice commands to reach a worldwide audience. Characteristics of the Community: establishing social features that allow users to exchange culinary techniques, recipes, and streaks, so creating a thriving home cooking community. Driven by our love for clever solutions that turn routine cooking into an interesting, effective, and enjoyable journey, FridgeAI keeps developing.

Built With

Share this project:

Updates