This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend
What I Built
Every single evening, the same universal question echoes through kitchens everywhere: "What on earth should I cook tomorrow?"
For this challenge, I built Tomorrow's Tadka 🍳—an intelligent, culinary-inspired web application designed specifically for my mom. Instead of staring blankly into a full fridge or wondering what to make with random leftovers, she now has a personal AI chef that solves two distinct everyday dilemmas:
🛒 Leftover Rescue Mode: Toss in a random mix of remaining ingredients (like leftover cooked rice, half an onion, and some curd), and the app crafts gourmet home-style recipes out of them.
✨ Seasonal & Weather Surprise Mode: Automatically detects local weather, seasonal vegetables, and regional context to suggest comforting, highly relevant meals.
Demo
Live Deployed App: Check out Tomorrow's Tadka on Render
The live demo runs OpenAI's open-weight gpt-oss-20b through Groq's free tier. The private, run-it-yourself mode uses Meta's Llama 3.2 3B locally through Ollama, so your ingredients never leave your computer.
💡 A Quick Note on the Live Demo:
Because this app is hosted on a free cloud tier, the server may "sleep" after a period of inactivity. If the first recipe generation takes 30–60 seconds to wake up, don't worry—the kitchen is just preheating!However, Tomorrow's Tadka is designed to live privately on your own machine. If you want an instant, lightning-fast experience that keeps your kitchen data 100% offline and secure, we warmly recommend following the setup steps in the GitHub repo to run it locally on Ollama!
Code
You can explore the full source code and repository structure over on GitHub:
👉 GitHub Repository - Tomorrow's Tadka
2005Mohammadali
/
moms-meal-planner
An intelligent, local AI-powered meal planner that transforms kitchen leftovers or leverages live weather, locality, and seasonal produce into delicious family recipes.
🍳 Tomorrow's Tadka
"The question we all ask every evening: What should I cook tomorrow?"
Tomorrow's Tadka is a chef-inspired web application designed to help you make the most of what you have in your kitchen or align your meals with the current weather and season. Powered locally by Llama (via Ollama or Groq), it delivers practical, home-style recipes with a warm culinary interface.
✨ Key Features
- 🛒 Leftover Rescue Mode: Input random fridge ingredients to instantly generate creative meals using basic pantry staples.
- ✨ Surprise Me Mode: Automatically detects your location, live weather, and seasonal produce to suggest comforting, highly relatable local dishes.
- 🔒 Privacy-First: Runs locally on your machine via Ollama, keeping your data private.
- 🎨 Chef-Inspired UI: Features custom culinary typography, an interactive card layout, and smooth mode-switching transitions.
🛠️ Tech Stack
- Frontend: HTML5, CSS3, Vanilla JavaScript (Custom culinary design system)
- Backend: Python, Flask, Gunicorn
- AI Integration:…
How I Built It
Tomorrow’s Tadka is built around an open-source, lightweight AI stack designed for speed and privacy:
- Open-Weight Model: At its core, the app runs on Llama 3.2 (3B) via Ollama, executing local inference straight on the machine.
- Full-Stack Architecture: A clean, responsive HTML/JS frontend (index.html) paired with a Python Flask backend (server.py) that parses structured JSON prompts into rich, card-based UI elements.
- Open APIs for Context: It integrates with Open-Meteo open weather and geocoding services to fetch live environmental data without relying on heavy commercial tracking tools.
Why Does Open Innovation Matter?
Open innovation was the absolute key to making this project practical for a real home kitchen:
- 100% Private & Local: Because it runs on Ollama, it can operate locally on a laptop with zero internet dependency. My mom’s random fridge contents and location data never get shipped off to a corporate server she doesn't control.
- Zero Cost to Run: There are no paid API tokens or subscription walls to worry about. It costs absolutely nothing to spin up an infinite number of dinner ideas.
- Total Flexibility: The open-weight approach made it trivial to swap model runners or tweak prompt behavior to match authentic home-style cooking rather than generic corporate AI responses.
💡 A Note on Hosting vs. Local Usage:
While we’ve set up a public deployment so you can easily test Tomorrow's Tadka right away, this app was truly built to live quietly and privately right on your own machine. If you fall in love with using it for your daily meal planning, we warmly recommend following the instructions in the GitHub repo to run it locally. Running it locally keeps your kitchen data completely private, costs nothing, and gives you the coziest, most seamless experience possible!
🎁 Handing It Over
When I first showed it to her, she was deeply skeptical—fully expecting another complicated app or gadget she’d never touch. But when she typed those leftover random items and the screen instantly spun up a practical, comforting recipe instead of making her scroll through a 2,000-word food blog essay or watch lots of video on YouTube just to find a spice ratio, her expression completely changed.
She didn't give a grand tech speech or clap. She just looked at the screen, nodded approvingly, and said, “Wait, it actually didn't ask me to order anything fancy or go to the market? Keep this tab pinned on my phone.”
For any developer, that’s the ultimate win 😊.
Prize Categories
Used Render to deploy my app.
Top comments (0)