The dilemma that began in front of the refrigerator

One day, while organizing the refrigerator, I threw away expired tofu and wilted vegetables. This was the third time this month alone. I had definitely bought them at the supermarket intending to cook, but whenever I opened the refrigerator, I couldn't think of what to make, so I ended up ordering delivery, and the ingredients were thrown away like that. 'What are some ways to reduce food waste in the refrigerator?' This question began to linger in my mind.

I threw that question at Claude. As we talked, an idea naturally came to mind. "What if there was an app where you took a picture of the ingredients in the refrigerator, and AI automatically recognized them and recommended a recipe?" Claude said it was a good idea and suggested we make it together. That is how K-Refrigerator Raid began.

Until the idea materialized

At first, it was a vague idea. I only had the big picture of an 'ingredient recognition app,' but I had absolutely no idea what specific functions were needed or how to build it. So, I started by writing a project plan with Claude. We organized, one by one, who the user would be and what the core features would be, as well as the user flow.

The first thing we did during the planning phase was research the Google Play Store. We checked to see if similar apps already existed. After searching, we found about 3 to 4 apps with similar concepts. However, they differed from what I had in mind. The existing apps focused on recipe search, and features such as automatic ingredient recognition, expiration date management, or grocery shopping integrated with Coupang were either non-existent or lacked. Analyzing the competing apps actually made our direction clearer.

There are three main reasons why we created this app.

It wasn't simply because the idea was good. There were three practical reasons why I decided to create this app myself.

The first reason — I wanted to personally solve the problem of food waste in the refrigerator.

I knew that the recurring food waste each month was not simply a matter of money, but a matter of habit. If you can immediately know 'what you can make with these ingredients' when you open the refrigerator, you can start cooking instead of turning on a delivery app. I believed that a single app could change a habit.

I felt that this app was particularly necessary for people living alone or the younger generation who are not accustomed to cooking. It could also be useful for those who want vegan meals or those who plan their weekly meals in advance. By saving meals and managing ingredients accordingly, you can reduce unnecessary grocery purchases and, consequently, lower your dependence on delivery food. With delivery fees rising these days, I thought that simply utilizing refrigerator ingredients properly could significantly reduce monthly food expenses.

The second reason — I wanted to experience the entire app development process firsthand.

I am currently developing other types of apps and game apps as well. To properly carry out these projects, I had to personally experience the entire process, from app development to Google Play registration. My thinking was that even if the app didn't work out, I would at least learn the whole process. There was no way to properly know how much it would cost, how long it would take, or what preparations were needed—except by doing it myself.

The third reason — I wanted to actually implement a monetization model.

I wanted to create a structure where I could offer the app for free but generate revenue through Google AdMob ads and Coupang Partners links. I needed experience with monetization on a different platform—an app—similar to AdSense for blogs. Knowing something in theory is a completely different story from actually implementing and operating it.

My honest feelings when I first started building it.

The moment the project plan was completed and I was about to start development in earnest, I was honestly half-doubting. My first thought was, "Will this even work?" I wasn't sure if it was realistically possible to create an app solely through conversation with AI, given that I had never learned coding and had absolutely no experience in app development.

However, at the same time, I also had the thought, "Let's give it a try even if it doesn't work out." I believed that even in the worst-case scenario, if there were no finished app, simply learning what I learned during the development process—how to structure an app, how to register on Google Play, and how to connect APIs—would be valuable enough. I decided to focus on what I gained from the process rather than the result.

Creating a Project Plan with Claude

Before starting development, I drafted a project plan with Claude. We organized, one by one, what the core features were, how users would use the app once they opened it, and how the revenue structure would work. As someone who didn't know coding, conversations with Claude served as both a planning tool and a development tool.

We combined the features I initially envisioned with those Claude suggested to create the final list of features. Automatic ingredient photo recognition, cooking recommendations, recipe saving, automatic calorie calculation, Coupang-linked grocery shopping, and expiration date management — these six features were confirmed as the core functions of K-Cold Food Clearance. Looking back now, this planning stage was the most crucial time. Development does not waver unless the direction is clear.

Once the planning was complete, the question "Where do I start?" grew larger than the doubt "Will this work?" The process of finding that answer begins in Part 3. If you are new to developing with AI tools, Truly Free! I recommend reading Comprehensive Guide to AI Tools You Can Use Right Now along with this.

Conclusion — The Idea Started with a Question

There are three main reasons why k-Naeng-tteol was created: solving the problem of refrigerator food waste, gaining experience with the entire app development process, and implementing a monetization model. It wasn't a grandiose business plan, but a single question stemming from an everyday inconvenience that led to a three-month development journey.

You can get started even if you don't know how to code, as long as you have an idea and can converse with AI. I learned through this process that simply starting is more important than perfect preparation. In Part 3, I will discuss how we actually started development and the process of creating the first version as a web app.