Inspiration
Sometimes you stand in front of your pantry with a few ingredients, no motivation, and no plan — but a flicker of curiosity. “I have garlic, pasta, and spinach… what can I do with this?” Pantry-fy is built for that exact moment — not to save the world, just to save dinner. It's like asking a quietly brilliant friend:
“Here’s what I’ve got... any ideas?” No pressure. No waste crusade. Just smart, gentle help from a kitchen companion that actually gets you.
What it does
- You type in your available ingredients
- Pantry-fy suggests real recipes you can make using them
- You can also save recipes you like or track what’s in your pantry
- It's simple, fast, and feels kind of like magic when it works (which it does)
How we built it
Frontend: React + Bootstrap for a clean, responsive experience Ingredient images come from Spoonacular’s API Recipe search powered by Spoonacular Pantry, user, and saved recipes stored in MongoDB
Challenges we ran into
Gemini AI integration was tricky — especially prompt handling and model responses Dealing with edge cases in ingredients (e.g., typos, uncommon foods) Recipe search + user pantry + AI recipe generation = balancing logic complexity Making it feel natural, not overwhelming
Accomplishments that we're proud of
We have a functioning webpage and several other pages (they are not connected yet).
What we learned
Prompt engineering and fine-tuning AI in real apps Handling real-world APIs and building robust fallbacks (like broken image handling) How to structure a small full-stack app that feels personal and functional INFP fact: food apps can be chill too — they don’t need to scream “optimize!”
What's next for Pantry-fy
🤝 Add friend sharing (like “here’s what I made with your ingredients”) 💾 Optional user login with cloud-synced pantries 🍳 Gemini-generated experimental recipes for creative cooking 📝 Smart shopping suggestions for what’s “almost possible” with current pantry 🌙 Dark mode. Because... obviously.
Log in or sign up for Devpost to join the conversation.