MoodMeal AI

AI recipe chatbot for mood, ingredients, weather, and daily cooking

MoodMeal AI is a Mongolian-focused recipe chatbot that recommends meals based on the user's mood, available ingredients, weather, meal time, and serving count.

It is designed to feel like a practical AI assistant for everyday home cooking, not just a generic chatbot.

---

Overview

MoodMeal AI is a recipe recommendation chatbot that helps users decide what to cook.

Instead of only asking for ingredients, it considers multiple real-life factors:

Input	Why It Matters
Mood	Food choices often depend on how the user feels
Ingredients	Recommendations should match what the user has
Weather	Cold, hot, rainy, or snowy weather can affect meal preference
Meal time	Breakfast, lunch, dinner, or snack changes the recommendation
Serving count	Recipe portions should match the number of people
Local context	Mongolian users may prefer familiar and practical meals

---

Why I Built It

I wanted to build an AI app that feels useful in daily life.

Many AI chatbots can generate recipes, but they often feel too generic. MoodMeal AI tries to make food recommendations more personal by using mood, ingredients, weather, and meal time together.

I also wanted to build something that can be useful for Mongolian users and home cooking.

This project helped me practice:

• AI API integration
• backend development with Flask
• SQLite data handling
• recipe recommendation logic
• prompt design
• user-focused product thinking

---

The Problem

Choosing what to cook can be surprisingly hard.

Users may think:

Problem	Example
I do not know what to cook	“Өнөөдөр юу хийх вэ?”
I only have a few ingredients	“Надад өндөг, гурил, мах байна”
I want food that matches my mood	“Ядарсан байна, амархан хоол хэрэгтэй”
Weather affects what I want	“Хүйтэн өдөр халуун хоол идмээр байна”
Normal recipe apps are too broad	Too many results, not enough personalization

MoodMeal AI tries to reduce this friction by acting like a small food decision assistant.

---

The Solution

MoodMeal AI recommends recipes using user context.

User input
    ↓
Mood + ingredients + weather + meal time + serving count
    ↓
Backend processing
    ↓
Recipe / ingredient data
    ↓
Gemini AI response
    ↓
Personalized meal recommendation

The goal is to make recipe suggestions feel more personal and practical.

---

Key Features

Feature	Description
Mood-based recommendation	Suggests meals based on how the user feels
Ingredient-aware suggestions	Uses available ingredients to guide recipes
Weather context	Adjusts recommendations based on weather
Meal-time awareness	Breakfast, lunch, dinner, or snack context
Serving count	Considers how many people the recipe is for
AI chatbot interface	User can ask naturally
SQLite storage	Stores or manages app data locally
Gemini AI integration	Generates personalized recipe responses
Mongolian user focus	Designed with local users and daily cooking in mind

---

Architecture

flowchart TD
    U[User] --> UI[Chat Interface]
    UI --> API[Flask Backend]

    API --> INPUT[Input Processor]
    INPUT --> MOOD[Mood Context]
    INPUT --> ING[Ingredient Context]
    INPUT --> WEATHER[Weather Context]
    INPUT --> TIME[Meal Time Context]
    INPUT --> SERVE[Serving Count]

    API --> DB[(SQLite Database)]
    API --> RECIPE[Recipe Data / TheMealDB]
    API --> PROMPT[Prompt Builder]

    MOOD --> PROMPT
    ING --> PROMPT
    WEATHER --> PROMPT
    TIME --> PROMPT
    SERVE --> PROMPT
    DB --> PROMPT
    RECIPE --> PROMPT

    PROMPT --> AI[Gemini AI]
    AI --> RESPONSE[Personalized Recipe Recommendation]
    RESPONSE --> UI
    UI --> U

Loading

---

Recommendation Logic

MoodMeal AI combines user context before generating a recommendation.

Context	Example
Mood	tired, happy, stressed, hungry
Ingredients	egg, flour, beef, rice, potato
Weather	cold, hot, rainy, snowy
Meal time	breakfast, lunch, dinner
Serving count	1 person, 2 people, family
Preference	easy, warm, healthy, quick

Example:

User:
I feel tired, it is cold outside, and I have beef, potato, and onion.

MoodMeal:
Suggests a warm, filling, easy dinner using those ingredients.

---

Example User Flow

1. User opens MoodMeal AI.
2. User enters mood and ingredients.
3. App collects meal time, serving count, and weather context.
4. Backend builds a structured prompt.
5. Gemini AI generates a recipe suggestion.
6. User receives a practical meal recommendation.

---

Tech Stack

Area	Tools
Backend	Python, Flask
Database	SQLite
AI	Gemini AI
Recipe Data	TheMealDB / recipe data
UI	HTML, CSS, JavaScript
Product Type	AI recipe chatbot
Target Users	Mongolian home cooking users

---

Project Structure

MoodMeal/
├── app.py
├── database/
│   └── moodmeal.db
├── templates/
│   └── index.html
├── static/
│   ├── style.css
│   └── script.js
├── data/
│   └── recipe / ingredient data
├── requirements.txt
└── README.md

Folder names may vary depending on the current version of the project.

---

Quick Start

Clone the repository:

git clone https://github.com/BeBecpp/MoodMeal.git
cd MoodMeal

Install dependencies:

pip install -r requirements.txt

Run the app:

python app.py

Open in browser:

http://127.0.0.1:5000

---

Environment Variables

Create a .env file if the project uses API keys:

GEMINI_API_KEY=your_api_key_here

Never commit real API keys to GitHub.

---

Example Prompts

User Input	Expected Output
“I am tired and want something warm”	Easy warm meal recommendation
“I have eggs, flour, and milk”	Recipe using available ingredients
“It is cold today”	Warm soup or filling meal idea
“I need dinner for 3 people”	Dinner recommendation with portions
“Надад мах, төмс, сонгино байна”	Mongolian-friendly recipe suggestion

---

What I Learned

While building MoodMeal AI, I learned how to turn a simple chatbot idea into a more useful product.

The hardest parts were:

• making recommendations feel personal
• combining multiple user inputs
• designing better prompts
• connecting backend logic with AI output
• thinking about local users
• making the app useful instead of just “AI for fun”

This project helped me understand that AI apps should not only generate text.
They should help users make decisions.

---

Current Limitations

Limitation	Future Fix
Recipe data can be limited	Add more local Mongolian recipes
AI output may vary	Add stronger response formatting
Weather context may be basic	Add real weather API integration
No user accounts yet	Add saved preferences and history
Limited nutrition info	Add calorie and nutrition estimates
UI can be improved	Add cleaner mobile-first design

---

Future Improvements

• Add more Mongolian recipes
• Add real-time weather API
• Save user favorite meals
• Add dietary preferences
• Add allergy warnings
• Add nutrition information
• Add weekly meal planning
• Add grocery list generation
• Add voice input
• Improve mobile UI
• Add multilingual support

---

Portfolio Summary

MoodMeal AI is a practical AI recipe chatbot focused on personalized daily cooking.

It demonstrates:

• Flask backend development
• SQLite data handling
• Gemini AI integration
• prompt engineering
• recommendation logic
• local user-focused product thinking
• AI app design for everyday use

This project represents my interest in building AI tools that are simple, useful, and connected to real daily problems.

AI should help people make better everyday decisions.