PantryPal

🧠 Inspiration Every year, over 1.3 billion tons of food is wasted globally—often due to forgotten items expiring in home kitchens. At the same time, millions face food insecurity. We wanted to build a smart, AI-powered pantry assistant that not only tracks food before it expires, but also recommends recipes and donation options to reduce waste and hunger.

🚀 What it does PantryPal is an intelligent web app that helps users:

Add food items via photo, barcode, or manual entry

Track expiry dates and receive timely reminders

Automatically generate recipes using AI based on what's about to expire

Find nearby food banks to donate surplus food

Visualize pantry usage with charts and analytics

Bonus features include:

Tavus-powered personalized video summaries for recipes

Hugging Face LLM for natural language recipe generation

chromaDB RAG to retrieve food facts and recipe suggestions

ElevenLabs (optional) for voice-based instructions or accessibility

🛠️ How we built it We built PantryPal entirely on Bolt.new, a React-based no-code/low-code platform. Key integrations include:

MongoDB Atlas (with RAG/Vector Search) for storing food items, expiry logs, and recipes

Hugging Face APIs for OCR and LLM-based recipe generation

Tavus to generate personalized recipe video summaries

Chart.js embedded in Bolt.new for dashboard analytics

Twilio for expiry notifications

Authentication with secure email/password & Google OAuth

The project is entirely front-end driven, using HTTP endpoints and workflow logic on Bolt.new.

🧗‍♀️ Challenges we ran into Integrating multiple external APIs (OCR, LLMs, Tavus) into a low-code environment required careful orchestration using Bolt.new’s HTTP actions

Handling OCR inconsistencies from poor-quality food labels

Designing the auto-recipe workflow to handle multiple scenarios (selected items vs. near-expiry auto-selection)

Keeping user experience clean while supporting both manual and AI-powered features

🏆 Accomplishments that we're proud of Built an end-to-end AI-powered pantry manager in just a few days!

Successfully integrated OCR, LLM, RAG, video generation, and analytics into a single Bolt.new app

Made it intuitive for everyday users to use AI without technical knowledge

Created a solution that solves real-world problems—food waste, hunger, and household efficiency

📚 What we learned How to rapidly prototype complex AI workflows using Bolt.new

How Retrieval-Augmented Generation (RAG) can enhance contextual AI output

The power of integrating AI with clear user interfaces—making technical tools feel human and helpful

The value of UX simplicity in social impact apps

🔮 What's next for PantryPal Mobile PWA support for faster scanning on the go

Integration with local grocery APIs for smart replenishment suggestions

Pantry sharing for families or flatmates

Voice assistant for accessibility (via ElevenLabs)

Reward points system for donations and food waste reduction

Launch in partnership with food banks and zero-waste nonprofits

Built With

• bolt.new
• github
• huggingface
• javascript
• langchain
• llm
• mongodb
• netlify
• rag
• react
• sqlite
• tavus