Recipe Lens: Visual Nutrition Intelligence
How QPOI engineered a "visual-first" React Native app that transforms food photography into instant recipes and nutritional data.
Duration
3 Months
Platforms
Mobile (iOS & Android)
Tech Stack
React Native (Expo), Gemini Flash
Engagement
MVP Engineering
The Challenge
Tracking nutrition and organizing recipes traditionally involves tedious manual input. Users struggle to estimate calories or identify dishes from photos without a scale or search engine.
The client needed a solution to:
- Identify Dishes: Instantly recognize food from raw images.
- Estimate Nutrition: Provide calorie and macro breakdowns without weighing.
- Remove Friction: Bridge the gap between a raw image and structured data.
QPOI's Role: Full-Stack Mobile Engineering
- Visual AI Integration
- Local-First Architecture
- Image Optimization Pipeline
- Subscription Management
Turning a Camera into a Nutritionist
We engineered a complete React Native application, focusing on image processing pipelines and offline resilience.
Instant Recipe Identification
Converts a photo of a finished dish into a structured recipe with ingredients, difficulty rating, and step-by-step instructions using Gemini Vision.
Visual Calorie Scanner
Analyzes food images to estimate calories and macronutrients (Protein, Carbs, Fat) with high accuracy, replacing manual entry.
Local-First "Smart Pantry"
Built a custom offline storage system. Images move from temporary cache to permanent local storage, ensuring users keep their data without internet.
Optimized Image Pipeline
Solved API latency issues by implementing an on-device image manipulation pipeline that resizes and compresses media before transmission, reducing latency by ~40%.
Mobile Engineering Fundamentals
CRUD Flows
Full create/read/delete functionality for the local recipe database and nutrition logs.
State Management
Custom hooks to manage complex subscription states and usage caps.
Compliance
Managed permissions for Camera/Photo Library and implemented compliant IAP restore flows.
Technology Stack
Building a Visual AI App?
We build the complex image processing pipelines that make simple user experiences possible.
Start a Conversation