GemEat

The inspiration for GemEat came from a common struggle: the "mental load" of maintaining a healthy diet while living a busy lifestyle. I realized that while many people want to eat better, the process of meal planning, tracking macros, and finding recipes is often too time-consuming. I wanted to use my background in Finance and my passion for Artificial Intelligence to create a tool that makes nutrition effortless and personalized through advanced AI.

​GemEat functions as a personal AI nutritionist in your pocket. It turns simple cravings or photos of food into shoppable recipes and personalized meal plans in seconds. By utilizing multimodal AI, it can analyze what you have in your fridge and generate instant, healthy meal ideas tailored to your specific health goals and palate.

How we built it

We built the application using a modern, scalable tech stack: ​Frontend: Developed with React Native and Expo to ensure a fast, responsive cross-platform experience. ​AI Engine: Powered by the Gemini API, handling complex tasks like multimodal image analysis and personalized nutrition coaching. ​Backend & Database: Integrated Supabase and Firebase for robust data management and user authentication. ​Monetization: Implemented subscription logic using RevenueCat

Challenges we ran into

One major challenge was ensuring the AI agent provided consistently accurate nutritional data while maintaining a natural, conversational tone. Perfecting the "agent chatkit" required significant iteration. Additionally, managing the deployment pipeline—from configuring environment variables in Netlify to navigating the submission requirements for the App Store and Google Play Store—presented a steep but rewarding learning curve.

Accomplishments that we're proud of

I am particularly proud of successfully integrating a sophisticated AI agent that can handle real-world tasks like meal planning from a photo. Transforming a concept from a simple idea into a fully branded app with its own logo and professional identity was a major milestone.

What we learned

This project was an incredible deep dive into AI integration and mobile entrepreneurship. I learned how to bridge the gap between complex AI logic and a user-friendly mobile interface. I also gained valuable experience in the full app lifecycle, including branding, marketing via social media, and implementing monetization frameworks.

What's next for GemEat

The next step is to expand GemEat's multi-agent capabilities, allowing for even more specialized assistants (like a "Fitness Agent" working alongside the "Nutrition Agent"). I also plan to refine the community aspects of the app, enabling users to share their AI-generated meal plans and success stories.

Built With

• adapty-?tools:-expo-router
• android-(google-play)-?cloud-services-&-backend:-supabase-(database-&-authentication)
• built-with-?languages:-javascript
• expo-?platforms:-ios-(app-store)
• for
• gemini-function-calling
• google-firebase-(google-ai-studio-integration)
• lucide
• netlify-(for-environment-variables/web-components)-?apis-&-ai:-gemini-api-(pro-&-flash-models)
• openai-agent-chatkit-(integrated-via-firebase)-?monetization:-revenuecat
• react
• sql-?frameworks:-react-native
• typescript