ValueMunch
A mobile friendly journey from pantry ingredients to meal ideas, guided cooking, impact tracking, and a simulated reward reveal.
1st Place · Build with TRAE · 31 January 2026
Problem and Context
People often have food at home but struggle to decide what to cook. We built ValueMunch to turn pantry photos or entered ingredients into meal ideas and guided cooking steps.
User Flow
- Capture or upload a pantry image, or enter available ingredients.
- Review recognized ingredients and generated recipe suggestions.
- Select a recipe and follow sequenced cooking steps and timers.
- Complete the meal to view impact information and a reward reveal.
- Use sample or cached recipes when AI processing is unavailable.
Capabilities
The expanded demo covers ingredient discovery, camera capture, recipe categories, guided cooking, profiles, photo storage, notifications, impact views, and simulated rewards. A separate pantry PWA focuses on three ingredients, an optional cuisine twist, generated recipes, and offline caching.
Architecture
The expanded application uses React, TypeScript, Vite, Tailwind CSS, Radix UI, Motion, Firebase Authentication, Firestore, and Storage. The separate pantry PWA design uses Next.js, Supabase, Vercel, Gemini and fallback model providers, IndexedDB, and a service worker.
Status
Public repositories contain source and planning documents for both variants. The former Firebase demo is retired, and the pantry PWA requirements remain a draft.
Team and Hackathon
Built by Richard Lee, Kanupriya, and Tanisha. We built ValueMunch in four hours at ByteDance using TRAE for planning, Google AI Studio for design experiments, and Firebase for deployment.