READING
0%
HENRYLIFEHACKS
← All guides
Guide · Interactive · 12 min

How I Built a Mac App Without Writing Code

I can't program. There is a real application running on my Mac right now that I made. Both of those sentences are true, and this guide explains how.

No course to buy. No no-code builder. Just plain English, honestly reported.

The bit people don't believe

Every line of code in my app was written by AI. I wrote none of it, and I still couldn't read most of it.

What I did instead was the part AI genuinely cannot do: know what the thing should be, notice when it was wrong, and say so clearly enough to act on.

That turns out to be the actual job. The typing was never the hard part.

0

Lines I wrote

Not one.

96

Days start to finish

Not an afternoon.

41

Rebuilds

Before the good version.

2,392

Lines AI wrote

Website plus Mac app.

It's a loop, not a magic sentence

Nobody types one perfect instruction and receives a finished app. You go round a cycle, and you sit in the middle of it — the only part that knows what "right" looks like.

1 · Picture it

Know what should exist and who it's for. This never comes from the machine.

2 · Say it plainly

Ordinary words. "I don't want to open Chrome first" is a complete brief.

3 · Let it build

The code, the structure, the fiddly parts. Minutes, not months.

4 · Try it yourself

Use it like a stranger would. This is where the real problems surface.

5 · Say what you saw

Not "it's broken" — where it stopped. Then go round again.

↻ Repeat

Forty-one times, in my case. That's the bit nobody photographs.

The skill nobody mentions

I can't read code. I can read symptoms — and it turns out that's the skill that matters.

My app froze twice: once in a web browser, once on the Mac. Three months apart, two different programming languages. I found both faults myself, without reading anything technical.

Here's the whole trick. Try both descriptions:

"It's broken"
"It freezes after the word is captured"

Both are true. Only one of them is useful. Tap the second one.

Both freezes turned out to be the same mistake in different clothes — the app waiting on a signal that never arrived. Spotting that pattern is taste, not training.

The honest timeline

Anyone selling you a finished app in an hour is showing you a demo. This is what it really looked like.

Day 1

A description became a working website. Rough, but real and online within two days.

Day 5

Browsers started fighting. iPhones refused to play audio, one browser blocked the microphone entirely.

Days 5–90

Rewritten from scratch. It grew from 26 spellings to 467, from one language to three.

Day 96

A native Mac app with its own icon — and every browser workaround deleted in a single day.

The most useful thing I did all project

I removed features. Four buttons and all the sound came out, and the app got noticeably better. Deleting beats fixing more often than you'd think.

Four things that mattered more than code

Name the friction

"I have to open Chrome first" is a better brief than any specification. Annoyance is a spec.

Say what you saw

Where it stopped, not that it stopped. Precision beats vocabulary every time.

Deleting beats fixing

My two best decisions removed things. Smaller was better.

Keep going after it works

Most people stop at the first version. The good one was the fortieth.

Five questions

Answer all five and find out whether you've actually got it.

Your turn

What would you build?

I'm not a developer. I had an idea that annoyed me enough to fix. If you've got one of those and you want to know how to actually get it made, write to me and I'll show you how I did it.

hello@speaktospell.com

A real reply from a real person. No course, no funnel.

The app in this guide is Speak to Spell — free, no sign-up. Say a word, see how it's spelled, copy it.