Nutrient Risk Profiler

A web application to check the risk profile of food items, inspired by Yuka.

Features

• Product Search: Search for food items by name.
• Risk Scoring: Custom scoring algorithm based on nutritional values (Calories, Sugar, Fat, etc.).
• Detailed Analysis: View breakdown of nutrients and ingredients.
• Premium Design: Modern, responsive UI with smooth animations.

Scoring Methodology

The Nutrient Risk Profiler calculates a health score from 0 (Poor) to 100 (Excellent) for each product. The algorithm is a simplified heuristic inspired by Nutri-Score, adapted to provide a clear risk assessment.

How it Works

The score starts at a Base of 100 and is adjusted based on the nutritional content per 100g:

1. Negative Factors (Penalties)

Points are subtracted for nutrients that should be limited:

• Energy (Calories): Linear penalty from 0 pts (300 kcal) to 40 pts (700 kcal).
• Sugars: Linear penalty from 0 pts (5g) to 70 pts (50g).
• Total Fat: Linear penalty from 0 pts (5g) to 30 pts (35g).
• Saturated Fat: Linear penalty from 0 pts (1g) to 50 pts (10g).
• Sodium: Linear penalty from 0 pts (0.2g) to 35 pts (2g).
• Net Carbs: Linear penalty from 0 pts (35g) to 15 pts (70g). Formula: Carbs - Fiber - Sugars to avoid double-counting sugars.
• Additives: 1 point per additive + additional penalties for high-risk additives (e.g., Nitrites, Artificial Colors). Max penalty 40 pts.

Note: If sugar content is high (> 30g), the Fruit/Vegetable bonus is disabled to prevent sugary snacks from scoring artificially high.

2. Positive Factors (Bonuses)

Points are added for beneficial nutrients:

• Fiber: Up to +10 points (> 8g).
• Protein: Up to +15 points (> 15g).
• Fruits, Vegetables, & Nuts: Up to +15 points (> 80%).

3. Final Score

The final score is calculated as Base (100) - Penalties + Bonuses, clamped between 0 and 100.

Score Range	Risk Level	Color
90 - 100	Excellent	Dark Green
70 - 89	Good	Green
40 - 69	Moderate	Yellow
20 - 39	Poor	Red
0 - 19	Bad	Dark Red

How to Run

1. Install Dependencies:

   npm install

2. Start Dev Server:

   npm run dev:worker

Available Scripts

Script	Description
npm run dev	Frontend-only dev server (port 5173)
npm run dev:worker	Full stack with Cloudflare Workers (port 8788)
npm run dev:fresh	Clear cache + start fresh dev server
npm run cache:clear	Clear server KV cache only
npm run build	Production build
npm run test	Run unit tests

OCR Backend

The Live Score Calculator uses Gemini 2.5 Flash for nutrition label OCR. This requires a Google AI API key.

Setup

1. Get API Key:

   • Google AI Studio

2. Add Secret to Cloudflare Pages:

   npx wrangler pages secret put GEMINI_API_KEY
   # Paste your API key when prompted

3. Deploy:

   npm run build && npx wrangler pages deploy dist --branch main

Local Testing

To test the OCR locally, create a .dev.vars file in the project root:

# .dev.vars (git-ignored)
GEMINI_API_KEY=your_api_key_here

Then run with the worker:

npm run dev:worker

How it Works

User captures image → /api/ocr → Gemini 2.5 Flash → JSON nutrients → UI

The API extracts nutrition values per 100g and returns them as structured JSON.

Tech Stack

• React 19
• Vite
• TailwindCSS
• Cloudflare Workers (API proxy)
• OpenFoodFacts API

API Rate Limits

OpenFoodFacts has rate limits to be aware of:

Endpoint	Limit	Usage
Product lookup (barcode)	100 req/min	Barcode scans, pasting codes
Search	10 req/min	Text search queries

Tip: Barcode scanning/pasting is preferred over text search for both accuracy and rate limits.