How I re-built my personal website in 2 weeks
A new portfolio built in a 2 weeks, 4 months of experiments with Claude, and what I learned about my own taste along the way.
Finally designed and built my new portfolio website at spoony.work.
My old site was on Framer and mostly just a list of case studies, which started to feel a bit too 2020 and incomplete (check here). This time I wanted it to feel personal I added a “play” section for experiments that I have built or shipped in the past year with random ideas, plus a “life” section to share more of who I am beyond work.
The details are the point
I built the whole thing inside Claude Code. Just describing what I wanted, and realizing along the way how much I care about the details. Most people stop once something works. I keep pushing until it actually feels right. The tone, the spacing, the composition, the flow, the tiny inconsistencies. All of it matters to me.
This time I wanted three rooms instead of one long scroll. Work, Play, Life. Three tabs in the topbar, treated as discrete spaces with their own rules. Before I touched a line of code I spent a few sessions just thinking about how those tabs should behave on mobile. The navigation that fits the size of a phone, with the right tap targets to switch between.
Some things surprised me with how easily they came together. I asked Claude to build me a small CMS so I could keep adding work without touching code, and it just worked, almost first try. A whole admin panel, image uploads, a project editor, all of it.
Other things you have to fight for. The contact form on Life is supposed to fold in 3D when you send it, like a real letter. The first version I tried had two sequential folds, horizontal then vertical, collapsing the page into a tiny envelope. Looked great in my head. In the browser the rotated flap kept leaking through during the rotation arc, because the outer container only clips the final size, not the path the flap traces through space. The paper sliced through itself in mid-air. I spent a whole night on it.
The version that ships uses a single horizontal fold. The bottom half is a real piece of paper with two faces, front and back, so the surface reads as the same letter folding onto itself. The send happens in parallel with the animation, so a slow API doesn’t pop the success state mid-fold and a fast one doesn’t cut the motion early. You can feel the difference.
What Play told me about myself
Play itself did something I didn’t expect. Looking at all the experiments lined up under one section, I started to see patterns I hadn’t named before. There’s a Czech salary comparison tool I’ve shipped jus because I keep finding small ways to make the metrics page look better. There’s an iOS boxing timer I built from scratch in SwiftUI because the apps on the App Store all looked wrong to me. There’s a CSS experiment about a single dot that arcs between filter pills, which exists for no reason except that I liked watching it move and learn how it was built by Petr Knoll. None of these were briefs. There's a Chrome extension I just shipped to the Web Store that lets you save any image on the internet to your own personal taste board.
Life section
Life is the room where I left all of my personal stuff.
An infinite canvas of my favorite photos I’ve taken, scattered across Mount Everest, Sapa in the clouds, Tokyo, a Thai boxing ring, a flight over New York in 2016. A toolbar I built where you can pick up a pencil and draw me a message. A tiny visitors counter that brings the nostalgia back. A contact form that’s a real cream-paper letter, and when you press send, it folds in 3D and lifts off the screen.
Each of those is an artifact. None of them are necessary in the resume sense. All of them say more about me than any case study. Photography taught me how things look. Design taught me how things work. The Vietnamese stamp belongs there because that’s where I am from.
What I learned from Claude
Self-initiated projects say more about a person than commissioned ones. There’s no client, no brief, no external motivation. You build them because you’re passionate about them, and the moment you finish, the only thing left on the page is the version of you who decided that this, specifically, was worth spending a weekend on.
The reason I could ship all of this in four months is that Claude does the building now. I describe a vision, look at what comes back, and push it again to my ideal.
It used to be how fast I could design. Now it’s how clearly I can describe the thing I want, and how honestly I can tell when it’s wrong.
That’s the part I didn’t expect. Working with Claude every day has made me a better designer, not because it’s faster, but because it’s forced me to articulate what “right” looks like every single time with all that comes into really designing a product not just as a canvas. You can’t fake taste with a model that only does what you ask. You actually have to know.
A portfolio is the epitome of that thinking. Still a work in progress, but it already feels a lot more personal because of all the small details I got to play with.