Pantry Pal: a recipe helper that knows what my friend can't eat

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

What I Built

Pantry Pal is a small web app that suggests a recipe from whatever ingredients you have at home, while avoiding the foods a specific person can't eat.

I built it for my friend Rahul, who is allergic to peanuts. Most recipes online don't take that into account, so he has to check every ingredient by hand before he cooks anything.

To use it, you type in the person's name, any foods they must avoid (for Rahul, peanuts), and what's in the kitchen. Pantry Pal then returns one simple recipe that takes under 30 minutes, with a short line confirming it follows those restrictions. The name and allergies are remembered in the browser, so they only need to be typed once.

Demo

Watch the demo here: https://drive.google.com/file/d/11AAUH9Q2Jrt7zD0SHBSBCxJ2BwjAdzdI/view?usp=sharing

Code

Karthikpillai77 / pantry-pal

How I Built It

I'm just starting out, so I kept the project small on purpose.

• Model: Gemma 3, an open-weight model from Google.
• Local inference: Ollama runs the model on my own computer.
• Backend: one Python file (app.py) using only the standard library. It serves the page and sends the prompt to Ollama.
• Frontend: one plain HTML file with a small form.

The prompt does the main work. It tells the model who it's cooking for, lists the foods to never use, and asks for one simple recipe plus a line confirming it's safe for those restrictions.

Why Does Open Innovation Matter?

Food allergies are personal health information. Because Gemma runs locally through Ollama, Rahul's allergies and what's in his kitchen never leave my computer. There's no account, no API key, and no data sent to a third party.

It also costs nothing to run and works offline once the model is downloaded. I can switch to a smaller or bigger model by changing one setting, which let me match the model to my laptop. With a closed API I'd be paying per request and sending health details to someone else's server.

One honest limit: a small local model can make mistakes, so this is a helper and not a medical tool. Anyone with a serious allergy should still check ingredient labels themselves.

Prize Categories

Best Use of Gemma: Gemma 3 runs locally through Ollama and is the entire AI core of Pantry Pal. It reads the person's allergies and the available ingredients, then writes a safe recipe, all on my own computer.