My Framer Site

Case study of a vibe coded AI application

A food delivery application where you can meet your health goals, even ordering in.

Problem Statement

In Indian cities, food delivery applications are popular. Health-conscious urban professionals, who live alone, are busy and cannot always make time to cook. Existing delivery apps are convenient, yet ingredients are not in the user’s control, and healthier alternatives are rigid or unaffordable. They know their nutrition goals but can not act on them.

Timeline

7th May - 29th May - 22 days

Team

2 UX Designers

My Role

Product requirement document, competitive analysis, customer journey mapping, paper wireframes, mid-fidelity prototypes, high-fidelity design (onboarding, product listing, order curation, cart, order profile, address), and development & deployment.

Tools Used

Claude

Figma Make

Chat GPT

Figma

VS Code

Git hub

Supabase

Netlify

Process

Empathize

AI Interviews

User Survey

Competitor Analysis

Define

User Persona

User Experience Map

Product Requirement

Document

MVP Feature list

Ideate

User Flow

Paper Wireframe Iterations

Mid Fidelity with AI

Typography & Color

High Fidelity Design

Test

Usability test on mid fidelity

vibe coded prototype with

Review of mid fidelity

prototypes

Prototype

Vibe coding

Setup database, git and

deployment

Research Insight

The research findings combine a quantitative survey of 33 health-conscious individuals (80% actively training) living in Bengaluru and 2 AI interviews.

85%

Low trust in healthy delivery claims

88%

In favor of auto nutrition logging

64%

Willing to pay ₹200 or more for a meal

67%

Goal driven - Building muscle or loosing fat

Feature Desirability

64%

Goal based suggestions

79%

Ingredient Customization

45%

Live kitchen camera

Manual nutrition tracking is too time consuming.

No time to meal prep or cook

Temptation is the biggest challenge.

Even after telling my cook to use less oil, after 2–3 days, she keeps putting more oil in my food

Competitor Analysis

User Persona

Priya

Senior Software Engineer

28 years old | Vegetarian

1.2L - Month

Goals & Motivation

Hit 110 g protein target and keep track of calories for muscle gain and fat loss.

Wants to know exactly what goes in her food.

Aspires to do Ironman 70.3 race.

Behaviors

Due to a time crunch, 4 meals are ordered from zoamto or swiggy per week for convenience and temptation.

Meal Prep does not work, especially with travelling to the office and so much training.

Pain Points

Too tired to cook food, personal cooks are not observant to nutrition, cravings makes me order from outside.

Unsure of nutritional claims of outside food, adulterated paneer, use of mayonnaise does more harm than good

Difficult to achieve high protein target with vegetarian meals.

Manual nutrition logging is exhausting.

User Journey Map

User Flow

Assumptions

The user will be interested in checking their daily calorie and macro intake in this application.

Curation of food may change taste and user will be comfortable with it.

User will be interested in checking live camera feed and the will have trust in the food supplied from this application.

User will be comfortable in ordering food from this application even though its main principle is not quick delivery.

Ideation & Design

Typography & Color

Regular

Medium

Semibold

#005C5B

#011C38

#5D6873

#E0F9FF

#AAB1B7

#ED7D3A

#EF2D56

#005C5B

#E0F9FF

#EF2D56

#011C38

#AAB1B7

#5D6873

#ED7D3A

#AAB1B7

#ED7D3A

#EF2D56

Onboarding

We began with paper wireframes for basic onboarding screens and used AI for a mid-fidelity design based on a color theme and material design. However, there were inconsistencies, so we created low-fidelity mockups to unify the onboarding steps.