It’s not a vibe until it’s shipped

How I avoided paying shipment costs and taxes by shipping my first app. The 2026 way to ship a beautiful app.

This article was going to be about vibe coding. But I really didn’t want to write about that. I still keep hearing and reading the same two main points that miss the mark:

• AI will make designers and/or developers redundant, it does everything and so well and blablabla
• It’s not possible to make a high quality product with AI, it’s so sloppy and never really designs the way it should

Sure, quite a few people made it to the “it’s a tool like any other” or “the photographer takes the photo, not the camera” conclusion. Still I see very little actually finished products with some honest numbers and reflections, so here we go.

I wanted a physical instrument and ended up with an app that kind of does the same but not exactly

Social media got me once again and showed me things, I didn’t know I wanted. A pocket sized chord synth from https://hichord.shop/ or however Tame Impala would call the Orchid from https://telepathicinstruments.com/.

For all non-music peeps: a save way to build songs that don’t sound like absolute turd.

The exact order of building a product with AI in 2026

You’ll soon enough realise not much has changed, everything you know about product still runs true, but it’s become amazing what on person can achieve with a little time and dedication.

The Problem

Let’s face it. I did not find it. It found me. I did not do much discovery, I assumed if there is so many successful and sold out products that there is a problem-market-fit. There was only one mistake to avoid. Confuse my problem —not wanting to pay so much money— with with the actual problem “I want to make music intuitively without worrying about theory or spending a long time mastering an instrument before writing anything decent.”

The Solution

Now, this is where things get interesting. Instead of ideating myself, I gave AI all the puzzle pieces I wanted to use.

• The original devices I liked.
• What I wanted to do with the app.
• How it should feel and behave, some basic principles.

No endless ideation sessions, mapping exercises, workshops, static designs in Figma. I let AI oneshot a mobile application. Had it installed on my phone. Did a few minor tweaks here and there.

What I achieved was a cheap clone of the hiChord instrument not nearly as exciting as what they advertised on social media. With underwhelming technological choices. Looking as if no one designed it.

I was soooo happy I had something working that for a little to long I tried to make it perfect. Soon enough I realised, I failed but luckily untrapped myself by iterating in the most radical way.

The Iterations

More often than not, in product, you’ll find reasons —read: excuses— to not throw something away and start over but instead refine, add, build on a shitty foundation.

I deleted everything and started from scratch. Scary? A bit. Exciting? A lot!

Only after having seen a version, I could tell what exactly I wanted and what I wanted to avoid. Now I could start with the core of the application and the core interaction that needed to be smooth, addictive and instant as fuck.

Building it step by step

Have you ever experienced agile work? I’m not talking about Scrum, SAFe, and all the theatre. Here is what I shipped one by one. Testable increments, each useful on its own. Each polish a deliberate improvement, not just a quest for perfectionism.

1. Core mechanic of triggering and altering a chord
2. Core Play screen with eight chords and readout what is playing at top
3. Apply a distinct branded design to it
4. Core Effects screen
5. Core Sequencer screen
6. Plenty improvements in the interactions, settings, parameter changes
7. Midi functionality and save

Followed by a bit of polish to make it ready to submit it to the app store.

The hard facts

With each new AI model release I see amazing videos and posts about one-shotted apps. Promising speed and fast success. Let’s put it into perspective and have an honest look at some figures.

In short: You start, then it takes way longer than imagined. If you have a subscription just for this plus the Apple developer account, it’s not quite cheap. Lastly, I haven’t earned a single penny yet; feel free to change that.

Support me

Try the app, show it to your musician friend, tell me if I should deep-dive on anything. We’re all still new to AI but I’ve rarely experienced such a steep growth within just a few months.

As much as I love building and designing, I really am no Marketer. Knowledge sharing is cool but now getting this out to people and at some point at least cover the costs … any help is appreciated!

It’s not a vibe until it’s shipped was originally published in Bootcamp on Medium, where people are continuing the conversation by highlighting and responding to this story.