A recipe app for university students in shared housing who want to eat according to their dietary goals.
Snap a photo of your fridge, select only your ingredients, and get recipes filtered by your diet and time constraints.
University students face daily friction when cooking:
- Open the fridge, see a mess of ingredients (some yours, some not), no idea what to make
- Want recipes that fit your diet but recipe apps don't filter this way
- Have 15 minutes before class and need something fast
- Half the fridge isn't yours, and no app accounts for shared living
1. Tap "Scan Your Fridge" (ideally you would take many if visibility is poor from 1 singular picture)
2. Take photo of ingredients
3. AI detects visible items
4. Select only what's yours (tap to include/exclude)
5. Choose your diet (high protein, keto, high calorie, vegetarian, etc.)
6. Set time available (10, 20, 30+ minutes)
7. Get matched recipes with nutrition info
8. View full recipe with step-by-step instructions
Fridge Photo Scanning Take a photo of your fridge, pantry, or counter. OpenAI Vision detects all visible ingredients and returns them with confidence scores.
Select What's Yours Detected ingredients appear as selectable chips. Tap to include only what belongs to you. Manually add items the AI missed.
Diet Filters
- High Protein: protein-focused meals
- Keto: low carb, high fat
- High Calorie: for bulking or weight gain
- Low Calorie: for cutting or weight loss
- Balanced: no specific restrictions
- Fun/Comfort: when you just want something good
Dietary Requirements Vegetarian, vegan, gluten-free, dairy-free, halal
Time Filters 10 minutes, 20 minutes, 30+ minutes
Equipment Constraints Microwave only, no oven, one pot, beginner friendly
Recipe Results Sorted by ingredient match, protein content, or cook time. Shows what you have vs what you'd need to buy.
| Layer | Technology |
|---|---|
| Frontend | React Native + Expo |
| Backend | Express.js |
| AI | OpenAI Vision API |
| Deployment | Vercel |
Photo -> Backend -> OpenAI Vision -> Ingredient list
-> User selects ingredients + filters
-> Backend matches recipes, scores by ingredient coverage
-> Sorted results returned to frontend
The OpenAI Vision prompt returns structured JSON:
[
{"name": "chicken breast", "confidence": 0.95, "category": "protein"},
{"name": "brown rice", "confidence": 0.90, "category": "carbs"},
{"name": "broccoli", "confidence": 0.85, "category": "vegetables"}
]Categories: protein, carbs, vegetables, fruits, dairy, condiment, spices, other
No other app combines:
- Photo scanning of fridge contents
- Ingredient selection for shared living
- Diet-specific filtering (keto, high protein, high cal, etc.)
- Time-based filtering
- Student-focused simplicity
The combination is the product.
University students who live in shared accommodation and want to cook meals that fit their specific dietary goals without the friction.