Skip to main content
DESKTOP VIEWPORT REQUIRED1024px+

"the admin is fu*ked up making the site responsive for mobile, please open the site on a desktop -thank you"

No desktop setup? Consider this alternative:

Progressive Web App · AI HealthNext.js 15 · Zustand · Gemini 2.5 Flash

AI-powered smart nutrition companion Progressive Web App (PWA) built with Next.js, Zustand, and Gemini 2.5 Flash for multi-modal meal analysis and streaming AI chat.

NutriLens AI PWA Interface dashboard

Client
NutriLens PWASmart Health Platform
Year
2025Mobile-First Web App
Role
Lead ArchitectFull-Stack & AI Builder
Stack
Next.js 15 · ZustandGemini 2.5 Flash · SSE
Status
LiveFully functional PWA

Calorie counting is traditionally slow and tedious, requiring manual ingredient search and weight estimations. NutriLens simplifies this by letting users photograph their food and instantly receive detailed nutritional logs.

Powered by the new Google GenAI SDK and the `gemini-2.5-flash` model, the app extracts ingredients, portion weights, calorie counts, and macro metrics directly from visual pixels.

It features a Streaming Context-Aware AI Chat utilizing Server-Sent Events (SSE). The AI acts as a personal nutritionist, referencing local meal logs and active targets to give hyper-targeted dietary advice.

Vision

Snap, track,
and optimize your health.

To eliminate manual calorie tracking bottlenecks. AI analysis combined with portion adjusters makes logging a meal take less than 10 seconds.

The app includes glassmorphic analytics cards and circular progress rings detailing daily macros, plus habit streak calendar heatmaps.

Product

PWA installer,
uncompromising offline state.

Zustand state partitions sync to local storage, keeping logged meals, calorie targets, and conversation history persistent even when the app is offline.

Portion scaling is managed via a live slider, letting users adjust serving sizes and immediately recalculate macro values client-side.

NUTRILENS system diagram overview
NUTRILENS full system architecture overview.
Telemetry

Interactive 3 Visual Calorie Tracker.

01Visual Plate AnalyzerDemonstrating the Visual Plate Analyzer functionality built into the core stack.Gemini Image Analysis
02Calorie Macro GaugesDemonstrating the Calorie Macro Gauges functionality built into the core stack.Zustand offline storage
03Streaming Nutritionist ChatDemonstrating the Streaming Nutritionist Chat functionality built into the core stack.SSE response pipelines
Architecture

Next.js 15 App
Gemini Multi-Modal API.

NutriLens leverages the speed of Turbopack and the Google GenAI SDK to process image data and stream chat messages safely.

PWA Client
  • Next.js 15
  • Zustand
  • Tailwind CSS
  • Recharts

Responsive mobile-first layout with Service Worker caching, dynamic macro circular gauges, and custom markdown message formatting.

Generative Backend
  • @google/genai
  • Server-Sent Events (SSE)

Server routes handling multi-modal analysis prompts and streaming nutritionist chat using Gemini's high-speed completion pipelines.

PWALauncher SupportAdd to Home Screen
SSEStreaming ChatReal-time responses
GeminiModel CoreMulti-modal vision analysis
Outcomes

Diet tracking,
made instant and visual.

10secMeal logging speedFrom snap to macro database write
Colophon

Engineered by byteWizard-zero.

Full-Stack & Generative Design
Zenith Soumya
AI PWA Developer
About Z.S

Zenith Soumya is an AI Developer and IoT system designer focused on high-performance human-interface web applications.

NutriLens runs locally as a PWA, storing session partitions in client-side state caches.

Live Demo ↗All workBack home