I rebuilt my oldest baby. Bye bye Elementor (not WordPress, WordPress you're the GOAT)

It’s August. It’s my birthday. I’m turning 39. I’m 39 now, actually, let me get that right the second time.

I spent a good chunk of this month doing two things: taking as much sun as Berlin is willing to hand out (not much, so you grab it while you can), and finally reworking my oldest baby, Un poco de sur.

If you’ve been around, you know the blog. It’s the thing that quietly started my whole career. It was built on WordPress from the very beginning, back when hosting didn’t come with a nice little button to install WordPress for you. You did it by hand. It’s not like that makes me some kind of hacker (it really doesn’t), but it’s funny to sit with how much time has passed. It started as a pre-designed theme I poked at with some HTML and CSS, exactly how I’d learned on every silly little blog before it, and it slowly grew into the last real version, a fully custom theme I shipped around 2020, in the middle of the pandemic.

And then, honestly? I kind of left it there.

I’ve written the odd article since, but somewhere in there I started enjoying travelling without needing to report back on it. The blog still brings in some revenue and still gets visits (it’s more alive than it has any right to be), but it stopped being my focus. I changed. A lot of what I’d write there now is really just personal reflection, the kind of thing I’ll probably translate and put here instead. And the travel itself lost its momentum, mostly because the way I travel now is fast, city-based, and honestly a bit repetitive.

But it was becoming a chore to keep alive. Plugins break, things need updating, something quietly stops working and eats an afternoon. And that math stopped making sense to me, because building with AI right now is FAST. I couldn’t justify pouring that kind of time into maintenance when everything else I build moves ten times quicker. I needed a way to keep working on the blog at the same speed as the rest of my projects, without it being a pain in the ass.

(A small casualty: Elementor. I used it to build the pages for years and I was a genuine fan, license and all, the ~200 euros a year kind of fan. This time I finally get to let it go. Subscription already cancelled. Bittersweet, but mostly sweet.)

The real problem was everything I refused to lose. Over 450 articles. Over 4,000 comments. A big part of whatever this blog is, and still does, lives in those two things. So I did my homework on headless CMSs (this is genuinely why I tried Keystatic on my personal site, and Tina, and the usual suspects I already use at work like Strapi). Nice tools. But none of them ever felt like THE replacement, and the comments were the part that scared me. Nobody had a clean answer for those.

And then it clicked: WordPress has a headless mode.

And here’s the honest part, because a lot of this comes down to hype, and to bias. The AI happily handed me version after version of CMS that could work, but it never really put the obvious one on the table: that WordPress can just stay WordPress, headless, if I want it to. To be fair, I had that fact buried somewhere in my own brain too, buried deep enough that I half-assumed it wasn’t even an option until I went and searched.

You’ve heard the story by now: everyone’s leaving WordPress because of AI. Honestly? I don’t buy it. WordPress is probably one of the biggest tools of all time, and it still quietly builds the majority of the websites out there. It’s not going anywhere.

If WordPress is losing anyone, I think it’s the people who never really needed it in the first place, the ones for whom it was always a bit of overkill. A simple landing page where the client just wants to swap some text? Sure, there are much better options for that now. But for a content-heavy site, hundreds of posts, thousands of comments, a decade of history? It’s still a killer, and genuinely hard to replace.

So that’s what I did. I kept WordPress, running headless in the background (on its own subdomain), reaching it through the REST API via Cloudflare so I can still change settings and edit properly. And I rebuilt the whole front end as a static Astro site, which, as a bonus, made it considerably faster.

First I replicated what I already had, the same design and the same functionality, because I didn’t want to lose the thing that worked. Then I let myself improve it, which with AI is suddenly easy and quick instead of a project in itself.

It was a lot of work, mostly the unglamorous kind: making sure every last bit of data came across safe and published. I also restructured the URLs (fixing a mistake I’d baked in years ago), which meant adding a couple of levels and a mountain of redirects. That gave me a small dip in organic traffic, nothing alarming, and I’m watching it daily to make sure it settles. So far, so good.

Which brings me to the part I didn’t expect to love: the blog is now on my daily list of projects to check. It had been left behind for so long, and if I’m honest I was a little afraid of letting it die. I don’t think I’m going back to travel blogging. But it still brings people in, it’s still very much alive, and it’s still a big part of who I am, even if it’s the dormant part.

So that was my August project. I’m genuinely proud of it. I’ve reworked the category pages, there’s a long list of things I still want to improve, and best of all it went back to being what it was supposed to be all along: a silly little fun project. Which is the whole point, isn’t it?