SnapNutrient is an AI-powered nutrition tracking app that helps users analyze their meals, receive personalized dietary suggestions, and engage with a health-conscious community through a built-in social platform.
Built with React, AWS (DynamoDB & S3), and OpenAI GPT-4o, SnapNutrient offers seamless image-based nutrient analysis and a rich user dashboard experience.
SnapNutrient uses few-shot prompting and state-of-the-art large language models (LLMs) to analyze user-submitted food images. Upon uploading a photo, the app generates a targeted prompt with embedded examples to guide the model in identifying food items and their macronutrient breakdown (calories, carbohydrates, proteins, fats). The result is a fast, detailed nutritional analysis with minimal user effort.
The SnapNutrient dashboard gives users immediate access to:
- Daily & weekly nutritional summaries
- Macronutrient and micronutrient breakdowns
- Filterable meal entries (breakfast, lunch, dinner, snacks)
- Real-time updates on edits or new entries
- Dynamic dietary suggestions powered by GPT-4o
Meal data is fetched from AWS DynamoDB (SnapNutrient_mealEntries), with associated images stored in S3. The dashboard ensures low-latency access through optimized batch queries and provides a smooth, responsive UI built with React.
Inspired by platforms like Instagram, the social feed allows users to:
- Share meal photos with captions
- Like and comment on posts
- Scroll through posts in a paginated, optimized feed
Posts are stored in DynamoDB (SnapNutrient_posts) with images referenced via S3 keys. The frontend uses lazy loading and batched S3 requests to minimize latency and improve performance.
Posting flow: Upload/take photo → Add caption → Preview → Share → Instant post display.
The goal of the social platform is to build a motivating, supportive user community — users land here first to inspire healthy habits through peer engagement.
- Frontend: React (Web & Mobile Responsive UI)
- Backend: AWS DynamoDB, AWS S3
- AI Integration: OpenAI GPT-4o for dietary recommendations
- DevOps: Vercel (for frontend deployment)
- Y. Huang
- S. Ji
- L. Long
- D. Wang
This repository is for demonstration and educational purposes. Please do not copy or reuse code without permission.
To restrict commercial use while allowing visibility, consider using the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) license, or view-only repo with a notice discouraging code reuse.
Have questions or feedback? Feel free to reach out via GitHub Issues or email: Langkunlong@gmail.com
This is a Next.js project bootstrapped with create-next-app.
#dependencies conflict issue when using npm, please try yarn install
npm install
# or
yarn install
# or
pnpm install
# or
bun installFirst, run the development server:
npm run dev
# or
yarn dev
# or
pnpm dev
# or
bun devOpen http://localhost:3000 with your browser to see the result.
This project uses next/font to automatically optimize and load Geist, a new font family for Vercel.
## Testing on mobile phone
First install ngrok to forward your port to outer network:
```bash
npm install -g ngrok
Start your app by running:
npm run devIn a new terminal, start ngrok:
ngrok http 3000ngrok will give you a link something like this:
https://d1ee-xx-xxx-xxx-xxx.ngrok-free.app Update your .env.local file:
NEXTAUTH_URL = https://d1ee-xx-xxx-xxx-xxx.ngrok-free.app Setup google oauth and add the link you created using ngrok. note: rememeber to change google client id and google secret if you're connecting to your own google oauth api
Example of what your google oauth should look like:
Authorized JavaScript origins
For use with requests from a browser
URIs 1
http://localhost:3000
URIs 2
https://25b5-69-166-116-174.ngrok-free.app
Authorized redirect URIs
For use with requests from a web server
URIs 1
https://d1ee-69-166-116-174.ngrok-free.app
URIs 2
http://localhost:3000
URIs 3
http://localhost:3000/api/auth/callback/google
URIs 4
https://25b5-69-166-116-174.ngrok-free.app/api/auth/callback/googleNow go to chrome and add the app to your homescreen, you're all set