I built a simple link locker 3 years ago. AI helped me turn it into a real product.

#ai #webdev #nextjs #sideprojects

The beginning

About 3 years ago I was learning frontend development. I wanted to build something real, so I made a simple link shortener with a password. Next.js + Firebase, nothing fancy. You create a short link, set a password, share it. That was it.

I deployed it and mostly forgot about it.

Surprisingly, people kept using it

Every now and then I'd check the analytics and see people were actually using it. Not a lot, but consistently. Some were sharing Wi-Fi passwords, some were sending credentials to coworkers. The core idea was simple enough that it just worked.

I always wanted to add more features but between work and life I never got around to it.

Then AI changed everything

A few months ago I started using Claude Code and my productivity went through the roof. Things that would have taken me weeks I was finishing in a day or two.

So I finally went back to my old side project and started building everything I'd been putting off:

Stripe payments for a Pro tier
E2E encrypted anonymous chat rooms
GPS location sharing
Secret memos
Polls, image sharing, audio sharing, boards, Q&A
12 language support
Blog with 120+ SEO posts
Chrome extension on the Chrome Web Store

What started as a simple link locker now supports 9 different content types. All password-protected, all with expiry settings. lock.pub

AI wasn't just for code

I used Claude Code for development, but AI helped with a lot more than that.

All the blog images were generated using Nano Banana. I have 120+ blog posts across 12 languages, and every single cover image was AI generated. Here's one of them:

Promotional graphics for the Chrome Web Store listing, social media, all the same. Stuff I would have either skipped entirely or spent days on was done in minutes.

Tech stack

Next.js 15 + React 19 + TypeScript
Firebase Firestore
CryptoJS for E2E encryption
Stripe for subscriptions
Tailwind CSS + Radix UI

What I learned

The biggest lesson was that a simple idea with real users is worth more than a complex idea with none. If I had tried to build all 9 content types from day one I never would have shipped anything.

AI didn't replace the thinking. I still had to decide what to build and how to structure it. But it removed the friction that kept me from actually doing it.