← Home

Project

Quackies

An AI-powered calorie tracker built around a simple idea — snap a photo of your meal and let AI do the logging. No food databases, no manual entry.


Why I built it

I wanted a different way to track calories without the bloat of existing apps. As a software engineer I could build it myself — it became the perfect project to practise an AI-native mindset, building genuinely around AI capabilities rather than bolting AI on as an afterthought.


Key features

Photo-based meal logging

Snap before/after eating, AI returns food names, portions, calories, and a confidence score. Up to 10 analyses per day.

Smart calorie budget

BMR via Harris-Benedict formula × activity PAL, adjusted for goal and pace. Live remaining figure: TDEE − eaten + burned.

Workout tracking

Log duration, estimates ~7 kcal/min, feeds directly into remaining calories.

Hydration tracking

Configure bottle size, log in fractions, visual progress on dashboard.

Duck mascot with 10 moods

Reacts to calorie status — happy on track, worried when under, grumpy when over.

Onboarding flow

6-step setup to compute a personalised daily calorie target before the user sees the dashboard.


Stack

FrameworkReact 19 + React Router 7 (Vite)
LanguageTypeScript
AnimationFramer Motion
Backend & AuthSupabase (Postgres, Auth, Storage)
AISupabase Edge Functions (image-based meal analysis)
PWAVite PWA Plugin

Still in progress

Three screens are actively being built:

LogFull meal history with day-by-day breakdown
ProgressWeekly summary view
ProfileEdit personal details and recalculate targets

Screenshots

Dashboard with live calorie remaining
Dashboard with live calorie remaining
AI meal analysis from a photo
AI meal analysis from a photo
Analyzed meal
Analyzed meal
Personalised setup flow
Personalised setup flow

View live app →View source →