Meme if no ingredients
Form to upload food images
Fetched ingredients from image
instructions of the recipe
Recommended recipes

FoodGasm: AI-Powered Recipe Suggestions from Available Ingredients

Inspiration

We created FoodGasm.AI to solve a real-world problem that many of us face daily: the struggle to decide what to cook with the ingredients we have on hand. As a team passionate about reducing food waste and promoting healthy eating, we wanted to develop a solution that would make meal planning easier, more efficient, and tailored to individual dietary needs.

What it does

FoodGasm is an AI-powered platform that suggests personalized recipes based on the ingredients available in a user’s kitchen. Users can either input ingredients manually or upload an image, and our system will recognize the items and generate recipe recommendations that align with their dietary preferences and nutritional needs.

How we built it

FoodGasm was developed using a combination of cutting-edge technologies:

Frontend: Built with React JS for an intuitive and responsive user experience.
Backend: Powered by FastAPI for high performance and seamless integration with AI models.
Image Processing: Utilized Microsoft Phi 3.5 LLM for accurate ingredient recognition.
Recipe Generation: Leveraged the Groq LLM for fast and precise personalized recipe suggestions.
Error Handling-memes: Implemented Memenome.ai to transform potential errors into engaging user experiences.

Challenges we ran into

Developing FoodGasm came with its fair share of challenges:

Ingredient Recognition: Ensuring accuracy across various lighting conditions and image qualities was a complex task.
Recipe Quality and Diversity: Fine-tuning the LLM to generate appealing, diverse recipes while maintaining dietary restrictions required multiple iterations.
Performance Optimization: Balancing AI computational demands with the need for a smooth user experience led us to creative optimization solutions.
Nutritional Accuracy: Creating a reliable nutritional analysis system for unique recipe combinations required extensive research and refinement.

Accomplishments that we're proud of

Successfully integrating AI-powered ingredient recognition and recipe generation into a seamless user experience.
Improving accuracy in identifying ingredients from images under varying conditions.
Optimizing AI models to generate diverse, personalized recipes that cater to different dietary needs.
Building a tool that can help reduce food waste while making cooking more accessible and enjoyable.

What we learned

Gained hands-on experience with computer vision and large language models in a real-world application.
Explored the complexities of nutritional analysis and dietary restrictions in AI-generated meal planning.
Strengthened our collaborative development skills, including version control, code review, and agile methodologies.

What's next for FoodGasm

We see immense potential for FoodGasm and plan to expand its capabilities further:

Enhanced Ingredient Recognition: Improving accuracy and expanding support for more ingredient types.
Advanced Dietary Customization: Allowing users to specify allergies, health goals, and preferred cuisines more precisely.
Smart Shopping List: Suggesting missing ingredients for recipes and integrating with grocery delivery services.
Community-Driven Recipes: Enabling users to share and rate recipes to improve recommendations.
Mobile App Development: Expanding FoodGasm into a dedicated mobile application for a more seamless experience.

FoodGasm is more than just an app—it's our step towards reducing food waste, promoting healthier eating, and making meal planning effortless for everyone!