System Requirements Document (SRD)

Project Name: earthy-app

1. Introduction

The earthy-app is a modern, AI-powered web and mobile application designed to generate personalized recipes based on ingredients provided by users. Developed with a focus on practicality and user convenience, the app empowers individuals to make the most of their available ingredients while exploring diverse cuisines and dietary preferences.

This document outlines the system requirements for the earthy-app, including functional and non-functional specifications, user personas, visual design elements, and technical considerations. The app will cater to users in Pakistan (PK) and beyond, ensuring a localized yet globally relevant experience.

2. System Overview

The earthy-app leverages cutting-edge AI technology to analyze user-provided ingredients and generate realistic, cookable recipes. The system will prioritize simplicity, accessibility, and personalization, offering features such as account functionality for cross-device recipe saving, dietary and cuisine filters, and smart suggestions for missing ingredients.

The app will be available on both web and mobile platforms, ensuring a seamless experience across devices. Its clean, minimal interface will make it easy for users to input ingredients, browse recipes, and save their favorites.

Key highlights of the system include:

• AI-powered recipe generation with detailed instructions.
• Support for multiple cuisines and dietary preferences.
• Cross-device recipe saving via user accounts.
• Smart behavior to suggest missing ingredients and avoid repetitive results.

3. Functional Requirements

Core Features:

• As a User, I should be able to input ingredients manually (comma-separated) or select from a list.
• As a User, I should be able to view multiple recipe options generated by the AI.
• As a User, I should see each recipe include:
  • Recipe name.
  • Ingredients list (with exact quantities).
  • Step-by-step cooking instructions.
  • Preparation time and cooking time.
  • Difficulty level (easy, medium, hard).
• As a User, I should see recipes prioritize using ONLY the ingredients I provide, with optional extras suggested if necessary.

Smart Behavior:

• As a User, I should receive suggestions for missing ingredients if my input is insufficient.
• As a User, I should be able to generate recipes in different cuisines/styles (e.g., healthy, desi, vegan, quick meals).
• As a User, I should not see repeated recipes for the same ingredients.
• As a User, I should be able to click a “Regenerate Recipe” button for new variations.

UI/UX:

• As a User, I should see a clean, minimal, premium interface.
• As a User, I should have an input field at the top for ingredients.
• As a User, I should see results displayed in a card format.
• As a User, I should be able to click a “View Full Recipe” button on each recipe card.
• As a User, I should be able to save recipes for future reference.

Advanced Features:

• As a User, I should be able to filter recipes by:
  • Cuisine type.
  • Dietary preference (vegan, keto, halal, etc.).
  • Cooking time.
• As a User, I should be able to click a “Generate Full Recipe” button to expand detailed steps.
• As a User, I should see loading states while the AI generates recipes.

Account Functionality:

• As a User, I should be able to register and log in to save recipes across devices.
• As a User, I should have a personalized dashboard to view my saved recipes.

4. User Personas

1. General User:

• Description: Everyday individuals looking for quick and easy recipes based on available ingredients.
• Goals: Save time, reduce food waste, and explore new recipes.
• Pain Points: Limited cooking knowledge, lack of time, and dietary restrictions.

2. Registered User:

• Description: Users with an account who want to save and access recipes across devices.
• Goals: Build a personalized recipe collection and access it anytime, anywhere.
• Pain Points: Losing recipes or not being able to sync across devices.

3. Admin:

• Description: System administrators managing the app’s backend, user accounts, and AI functionality.
• Goals: Ensure smooth operation, monitor user activity, and improve AI performance.
• Pain Points: System downtime, user complaints, and AI inaccuracies.

5. Visuals Colors and Theme

The earthy-app will feature a modern, nature-inspired color palette to evoke freshness and simplicity.

• Background: #F7F5E9 (Soft Cream)
• Surface: #FFFFFF (Pure White)
• Text: #2C3E50 (Deep Charcoal)
• Accent: #27AE60 (Fresh Green)
• Muted Tones: #BDC3C7 (Light Gray)

6. Signature Design Concept

The earthy-app homepage will feature an interactive ingredient bowl as its centerpiece.

Concept Details:

• Visual: A 3D-rendered bowl at the center of the screen, surrounded by floating ingredient icons (e.g., tomatoes, onions, spices).
• Interaction: Users can drag and drop ingredients into the bowl, which triggers a subtle animation (e.g., ingredients blending together).
• Animation: Once ingredients are added, the bowl “spins” briefly, and recipe cards appear below it in a cascading animation.
• Color Shifts: The background subtly changes hues based on the type of cuisine selected (e.g., green for vegan, orange for desi).
• Micro-interactions: Hovering over recipe cards causes them to “pop” slightly, and clicking a card flips it to reveal more details.

This design will make the app feel dynamic, engaging, and memorable, ensuring users enjoy every interaction.

7. Non-Functional Requirements

• Performance: The app should generate recipes within 3-5 seconds.
• Scalability: The system must support up to 100,000 concurrent users.
• Security: User data (e.g., saved recipes, login credentials) must be encrypted.
• Localization: The app should support English and Urdu languages.
• Accessibility: Ensure WCAG 2.1 compliance for users with disabilities.

8. Tech Stack

Frontend:

• React for Web.
• React Native for Mobile App.

Backend:

• Python with FastAPI.

Database:

• MySQL for structured data (e.g., user accounts, saved recipes).
• MongoDB for unstructured data (e.g., recipe generation logs).

AI Models:

• GPT 5.4 for user-friendly recipe generation.
• Claude 4.6 Opas for coding and academic logic.

AI Tools:

• Langchain for AI workflow orchestration.
• Litellm for LLM routing.

Orchestration:

• Docker for local development.
• Kubernetes for server-side orchestration.

9. Assumptions and Constraints

Assumptions:

• Users will have internet access to use the app.
• The AI model will be pre-trained and fine-tuned for recipe generation.
• Users will primarily access the app via mobile devices.

Constraints:

• Limited to ingredients and cuisines supported by the AI model.
• Initial launch will support English and Urdu only.

10. Glossary

• AI (Artificial Intelligence): Technology used to analyze ingredients and generate recipes.
• WCAG (Web Content Accessibility Guidelines): Standards for making web content accessible to people with disabilities.
• LLM (Large Language Model): AI models like GPT used for natural language processing.
• Frontend: The user-facing part of the app.
• Backend: The server-side logic and database management.