Flask web app that uses Meta Llama 4 Maverick 17B Vision (via IBM Watsonx) to analyze food images and provide detailed calorie counts, nutritional breakdowns, and personalized dietary coaching. Upload a food photo and get an instant AI-powered nutritional analysis.
Domain: Multimodal AI — Nutrition & Health
LLM: meta-llama/llama-4-maverick-17b-128e-instruct-fp8
Platform: IBM Watsonx.ai
| Component | Technology |
|---|---|
| Vision LLM | Llama 4 Maverick 17B (IBM Watsonx) |
| Backend | Flask |
| Image Processing | PIL + Base64 encoding |
| Frontend | HTML/CSS templates |
- Food image upload → Base64 encoding → Llama Vision analysis
- Calorie estimation per food item with portion sizes
- Macronutrient breakdown (protein, carbs, fats)
- HTML-formatted responses with bold section headers
- Flask session management
Multimodal AI (vision + text) · IBM Watsonx · Flask web development · Base64 image encoding · Nutritional AI
