I Built an App With Claude Code

I Always Thought Programming Was Beyond Me

Let me confess something. For the longest time, I believed programming was a world that had nothing to do with me.

Back in school, I had a brief encounter with programming in a class. I stared at the lines of English characters on the screen and my mind went blank. I couldn't tell what anything meant. Forget one semicolon and the whole thing breaks. The error messages? Complete gibberish.

That experience became a kind of trauma. "Programming = something I can't do" became an equation burned into my brain.

But after I started using AI, I kept coming across articles about "AI that writes code for you." At first, I thought, "There's no way it's that simple." But the more I looked into it, the more I kept hearing about a tool called Claude Code.

Discovering Claude Code

Claude Code is an AI coding assistant from Anthropic. In simple terms, it's a tool where "you describe what you want to build in plain language, and AI writes the code."

When I found out about it, two emotions were at war inside me.

One was hope. If I could really build an app just by giving instructions in my own words, maybe I could finally get past my programming trauma.

The other was fear. Maybe I'd just prove to myself again that "I can't do this."

In the end, the spirit of this blog -- "just try it" -- won out. If it doesn't work, I'll just write about that experience instead. That's what I told myself as I fired up Claude Code.

The First App -- What Should I Build?

The first challenge was deciding what to make.

Every beginner guide out there recommends a to-do list as a first project. But I wanted something different. Something I'd actually want to use.

After thinking it over, I settled on a simple notes app. Type text, save it. That's the only feature. Nothing flashy, but I wanted something I'd use in my daily life.

Putting "What I Want to Build" Into Words

I sat down with Claude Code and started describing what I had in mind.

My first instruction went something like this: "Build a simple notes app. I just need the ability to type text and save it."

Claude Code started generating code almost immediately. Lines of code streamed across the screen. Honestly, I understood almost none of it. But I could tell that AI was trying to build something.

A few minutes later, a prototype of a notes app appeared before my eyes.

A text input area, a save button, and a list showing saved notes. What I had described in words had actually materialized right in front of me.

My hands were shaking. I'm not being dramatic -- they genuinely were.

The Trial-and-Error Begins

Of course, everything wasn't perfect from the start.

The design, honestly, was pretty bland. The features were bare minimum. But it worked. I could type text, hit save, and the note was stored. That alone was enough to move me.

From there, I started iterating, one improvement at a time.

"Can you make the design a bit nicer?" I asked Claude Code, and the appearance cleaned up. "Add a delete function for notes" and a delete button appeared. "I want a search feature" and a search bar materialized.

Just by describing what I wanted in words, the app kept evolving. This experience fundamentally changed how I thought about programming.

Hitting a Wall

Of course, not everything went smoothly.

Several times, things didn't work the way I expected. I'd give AI instructions and get something different from what I had in mind. At first, I was frustrated: "Why doesn't it understand what I want?"

But then I realized something crucial. The problem was with how I was giving instructions.

A vague request like "make it prettier" doesn't tell AI what to do. But "set the background to white, increase the font size, and add more padding" gets you something much closer to your vision.

This is the same prompting lesson I wrote about in a previous article. The skill of giving AI clear, specific instructions turned out to be the most important skill in coding too.

Battling Errors -- With AI as Your Ally

As development progressed, I encountered errors. Over and over. Every time an error message appeared on screen, my school-day trauma came flooding back.

But this time was different. When you get an error, you just tell Claude Code about it.

"I got this error. Can you fix it?" Share the error message, and AI analyzes the cause and makes the fix. Sometimes it even adds an explanation: "This part was the issue. Here's what I changed."

Errors stopped being scary. That was a revolutionary shift for me. Errors weren't "failures" anymore -- they were "opportunities to improve." Because every time an error got resolved, the app was measurably better than before.

The Moment It Was Done

After working on it over several days, the notes app was finally complete.

Text input and saving, deletion, search. Simple, but the thing I'd imagined -- "I want something like this" -- was right there in front of me. And a person with essentially zero programming knowledge had built it.

I sat there staring at the screen for a while, lost in thought.

"I actually did it."

This might have been the most powerful moment of accomplishment since I started using AI. For over ten years, I'd convinced myself that programming was beyond me. Claude Code shattered that belief.

What I Learned -- The Importance of Understanding Code

There's one thing I want to be honest about. I don't fully understand the code Claude Code generated.

I built something that works. But if someone asked me to explain exactly why each line of code does what it does, there are many parts I couldn't. I see that as something to work on.

Rather than relying entirely on AI, I need to gradually read and understand the generated code. I've started studying the basics, bit by bit.

You don't need complete understanding before you begin. But I want to hold onto the mindset of "learning while building."

Advice for Fellow Non-Programmers

To close, a few words for anyone who shares my apprehension about programming.

Start with something small. Don't try to build something grand right away. A notepad, a counter, anything. Stacking small wins -- "this actually works!" -- is the most important thing.

Don't chase perfection. Even if the first version is rough, the fact that it runs is what counts. Design can be fixed later. Features can be added later. Focus on building "something that works" first.

Don't be afraid of errors. In the age of AI, errors aren't something you have to fight alone. Get an error? Ask AI. That alone will move most problems toward a solution.

Programming is no longer reserved for a special few. With AI as your ultimate partner, even beginners can turn their ideas into reality.

Next time, I want to write about how to publish the app I built to the internet. "I built it, but how do I actually get it out there?" I'll answer that question from my own experience.