Four months ago I decided to build a simple utility website.
The original plan was maybe 15–20 tools.
Instead, I somehow ended up shipping 120+ tools covering PDFs, images, developers, students, finance, AI utilities, and a lot of niche calculators.
Looking back, I made almost every mistake a solo developer can make.
Hopefully this saves someone else a few months.
**1. Stop waiting for the "perfect launch" **
Huge mistake.
Nobody cares how many features you have.
People care whether you solve their problem.
If I started again, I'd launch after 10–20 tools and let users decide what to build next. **2. Boring tools get used more than clever ones **
Instead, the most used ones were things like:
Simple wins.
People search for solutions, not engineering.
**3. Niche tools are an SEO superpower **
Almost nobody builds things like:
The competition is almost zero.
Those pages have become some of my favourites because they solve real problems.
**4. Browser APIs are incredibly underrated **
Compression.
Image editing.
QR generation.
JSON formatting.
Regex testing.
UUID generation.
Password generation.
CSV editing.
All of these can happen directly inside the browser.
That means:
**5. Docker can save you from cold starts **
My AI background remover worked perfectly locally.
Production?
Every first request timed out.
The reason:
The U2Net model (used by rembg) downloaded on first request.
Around 170 MB.
The fix?
Download the model during the Docker build instead.
One tiny change.
Massive improvement.
**6. Users hate unnecessary accounts **
No email.
No registration.
No "Create an account to continue."
People just use the tool and leave.
That's exactly what most utility websites should encourage.
**7. Performance matters more than animations **
Users never mentioned them.
They did notice:
Speed is a feature.
**8. Mobile users deserve first-class treatment **
If your utility only works nicely on desktop... You've already lost a huge percentage of users.
Every tool I build now starts with mobile in mind.
**9. Build things you personally need **
Open it on your desktop.
Scan a QR code.
Start typing on your phone.
The text appears instantly.
No login.
No app.
No emailing yourself.
It wasn't built because someone requested it.
It solved a problem I had every week.
Those are often the best products.
**10. Marketing is part of building **
It doesn't.
If nobody knows your project exists... It doesn't matter how clean your architecture is.
Building is comfortable.
Marketing is uncomfortable.
Unfortunately, both are required.
Where the project is today
The project has grown into DukoTools, a collection of 120+ free online tools.
Everything is free.
No signup.
Many tools work entirely in the browser, and the site can even be installed as a Progressive Web App for offline use.
I'm now focusing on improving the existing tools based on real user feedback instead of endlessly adding new ones.
If you have a favourite online tool that doesn't exist—or one that could be significantly better—I'd genuinely love to hear about it. Building is fun.
Building things people actually use is even better.