Cooks' Match

Matches purchased ingredients on a receipt with recipes that you can cook

Comment

NOTE: the name was changed in between when the video was recorded and when we came up with a new name for this.

Inspiration

We wanted to solve a common everyday problem: figuring out what to cook with the ingredients you already have. Too often, people either waste food or default to takeout because they don't know what meals they can make. Our inspiration came from the idea of reducing food waste while making cooking more accessible and fun.

What it does

Cooks' Match takes a list of ingredients you have on hand and generates recipes you can make with them.

  • You upload your pantry or fridge contents.
  • The app analyzes your ingredients.
  • It recommends recipes you can cook right now, plus near-miss recipes that just need one or two extra items.

This helps you save money, reduce waste, and make meal planning easier.

How we built it

We built the backend using FastAPI and MongoDB to handle user data and recipe storage. The recipe generation logic matches user-provided ingredients against a recipe database, prioritizing recipes that use the most available ingredients.

For the frontend, we designed a simple and clean interface via React to upload ingredients and view recipe suggestions.

Key technologies include:

  • OCR – used for reading recipe information from pictures and handling backend data processing
  • Gemini AI (Google Generative AI) – powers recipe intelligence, recommendations, and natural language
  • FastAPI for the API layer
  • MongoDB for storing recipes and ingredient data
  • Python for data processing and recipe matching
  • React (or vanilla HTML/JS) for the user interface (depending on your actual setup)

Challenges we ran into

  • Ingredient matching: Normalizing messy ingredient inputs like "2 cups flour" vs "flour" was tricky.
  • Data accuracy: Finding a reliable and complete recipe dataset was harder than expected.
  • Integration issues: Getting the backend and frontend to work smoothly together required debugging.
  • Deployment hurdles: Setting up and running the project locally with multiple services (MongoDB, FastAPI) was initially challenging.

Accomplishments that we're proud of

  • Built a working prototype that can match ingredients to recipes.
  • Solved the challenge of ingredient normalization to improve search accuracy.
  • Created a clean and intuitive interface for users.
  • Developed a project that could genuinely reduce food waste and help people cook at home.

What we learned

  • How to design and integrate a full-stack application using FastAPI, MongoDB and React.
  • The importance of good data hygiene when dealing with user-generated inputs.
  • How challenging it can be to make an algorithm feel "smart" with imperfect real-world data.
  • Effective teamwork and communication when building a project under time constraints.

What's next for Cooks' Match

  • Smarter recommendations: Incorporate machine learning to better suggest recipes based on cooking habits.
  • Recipe personalization: Allow filtering by dietary needs, preferences, or cooking time.
  • Mobile-friendly experience: Create a native mobile app or fully responsive web design.
  • Public recipe sharing: Let users share their favorite recipes with others.

Built With

Share this project:

Updates