MealMate: From Food Cravings to Shopping Carts

Contributors: Jonas Li, Jasper Liu, Mats Martinussen, Emil Klovning, Jason Ji, Qinghe Zheng

Affiliation: UC Berkeley

MealMate is an intelligent meal planning assistant that simplifies food shopping by seamlessly integrating recipe generation, ingredient availability checks, and personalized substitutions. Powered by advanced multi-agent collaboration and large language models (LLMs), MealMate delivers tailored shopping lists based on user preferences and real-time store inventory.

Quick access

• Demo video
• Project report

---

Key Features

• Chef De Cuisine Agent: Generates recipes, including detailed ingredients and instructions, based on user prompts.
• Line Cook Agent: Verifies stock availability in nearby stores, making smart substitutions based on user preference when needed.

---

Workflow Overview

Customer Interaction

A user inputs their desired dish (e.g., "I want pasta with tomato sauce tonight"). The Chef De Cuisine Agent generates multiple recipe options based on this request.

Recipe Breakdown

The Chef De Cuisine Agent provides the recipe details, breaking it down into ingredients with specified quantities.

Ingredient Availability Check

The Line Cook Agent cross-references the recipe ingredients with real-time store inventory:

• Identifies available items.
• Suggests substitutions for missing ingredients using the Recommender Agent.
• Returns a finalized, actionable shopping list.

Final Output

Users receive a tailored recipe and shopping list, optimized for their preferences, including:

• Selected or substituted ingredients.
• Pricing and location details.

---

Agent Configurations

Chef De Cuisine Agent

A specialized agent for generating recipes based on user requirements.

• Input & Configurations:
  • System Message: "You are an experienced chef. Break down the recipe based on user needs."
  • Example Prompt: "I want pasta with tomato sauce tonight."
• Process:
  • Leverages a database of valid ingredients.
  • Outputs ingredients and quantities.
    

Line Cook Agent

Responsible for verifying stock availability and providing substitutions and providing recommendations for substitutions, alternative dishes, or stores.

• Input & Configurations:
  • System Message: "Check ingredient stock in the database based on the recipe."
• Process:
  • Real-time stock data validation.
  • Provides recommendations for substitutions, alternative dishes, or store based on user preferences (e.g., price sensitivity, allergies).
• Output:
  • A shopping list tailored to the recipe and store availability.

---

Setup Instructions

Clone the repository:

git clone [email protected]:LIYunzhe1408/MealMate.git
cd MealMate

Backend Setup

1. Navigate to the backend directory:

   cd backend

2. Create and activate a virtual environment:

   python -m venv venv
   source venv/bin/activate   # On Windows: venv\Scripts\activate

3. Install dependencies:

   pip install -r requirements.txt

4. Launch backend:

   python app.py

Frontend Setup

1. Navigate to the frontend directory:

   cd frontend

2. Install dependencies:

   npm install

3. Launch the development server:

   npm start

Additional Notes

• Ensure you have Python 3.8+ and Node.js 14+ installed for compatibility.
• For deployment instructions, refer to the DEPLOYMENT.md file (if applicable).