DEV Community

Arjun Khanduri
Arjun Khanduri

Posted on

Build for a Friend: Fridgeful, an AI Kitchen Companion Against Food Waste

#devchallenge #weekendchallenge #hf26challenge

Hacktoberfest Weekend Challenge: Build for a Friend Submission 🤝

This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend

What I Built

I built Fridgeful (Inventory Mate) for my girlfriend who has recently moved to the UK and constantly ends up with groceries going bad in her refrigerator. Like many of us living busy lives, they would buy fresh produce with good intentions only to find forgotten vegetables turning into rotten mass.

Fridgeful solves this by acting as an intelligent kitchen companion:

  1. Live Pantry & Expiry Tracking: Visual cues and proactive expiration badges ("Today", "Tomorrow", "In 2 days") keep freshness front and center.
  2. Waste-Prevention Nudges: Smart browser notifications alert them before produce spoils.
  3. AI Recipe Studio: Instead of searching through recipes that require 15 new ingredients, one tap generates curated meal ideas built directly around what is already in their fridge prioritizing ingredients closest to expiry.
  4. Pantry Coverage Breakdown: Shows exactly what percentage of ingredients they already have, what are pantry staples (salt, oil, garlic), and what few items are missing.
  5. Aesthetics & Usability: Fluid micro-interactions powered by Framer Motion and offline PWA capability.

Demo

Code

You can view the full source code on GitHub:

GitHub logo arjun-khanduri / inventory-mate

Fridgeful

Application live at: https://inventory-mate.arkh18cs.workers.dev/

Fridgeful is a minimalist, high-performance inventory management tool designed to help you reduce food waste by keeping a clear view of your kitchen staples, tracking expiry dates, and generating personalized meal ideas from what you already own.

Features

  • Inventory Tracking: Easily add, track, and manage items in your fridge.
  • Expiry Monitoring: Visual cues for items nearing their expiration dates.
  • Smart Reminders: Browser notifications for expiring food.
  • AI-Powered Recipes: Generate meal ideas based on your current inventory using Groq's open-weight models, facilitated by a serverless Cloudflare Worker.
  • Save & Curate: Save your favorite generated recipes for later.

Tech Stack

Architecture

Fridgeful uses a static-export approach for the frontend, optimized for deployment on platforms like Cloudflare Pages. The AI recipe generation is…

View on GitHub

How I Built It

Fridgeful is engineered with a static, edge-optimized architecture:

  • Frontend: Built with Next.js 15 (App Router) and TypeScript, configured for static export (output: "export") for lightning-fast loads.
  • Styling & Fluid Motion: Styled using modern Tailwind CSS, dynamic dark/light mode via next-themes, and smooth UI state transitions (sliding active tab indicators, animated pantry lists, and spring modals) with Framer Motion.
  • Offline & PWA: Service worker integration enabling instant launch and offline access to existing fridge data via localStorage.
  • Serverless AI Proxy: A lightweight Cloudflare Worker acts as a secure, low-latency bridge between the client and Groq's high-speed inference engine, protecting credentials while serving CORS-compliant responses.
  • Open-Source AI Model: Powered by the open-weight Qwen 3.8 27B (qwen/qwen3.8-27b) hosted on Groq. We leverage Groq's ultra-low latency LPU inference with strict structured JSON output schemas to classify ingredients into have, staple, and missing statuses alongside clear, numbered cooking steps.

Why Does Open Innovation Matter?

Open innovation is what made building Fridgeful possible as an independent developer:

  1. Freedom of Choice: Using open-weight models like Qwen via open APIs avoids walled-garden lock-in. If we want to switch between Qwen, Llama 3, or Mistral, or run inference locally with Ollama, the interface remains modular and flexible.
  2. Speed & Accessibility: Pair open-weight intelligence with high-throughput engines like Groq, and recipe generation goes from a roughly 10-second wait to sub-second streaming. This turns an AI feature from a novelty into a real daily utility.
  3. Privacy & Simplicity: Open architectures allow simple, stateless serverless edge workers to handle user prompts without storing or tracking intimate household consumption habits.

Top comments (0)