People keep saying WordPress needs more. More features, more blocks, more flexibility baked into Core. And maybe they’re right — but I’ve started to think the real gap isn’t features. It’s connectors.
Someone has to look at what’s already there and say: wait, these pieces fit together.
(Not these connectors. Those are a different and genuinely exciting thing. I mean the human kind.)
That’s what this plugin is. Not a new idea. Not a framework. Not a page builder or a proprietary template or a settings panel with seventeen options you’ll configure once and never think about again. It’s two things Core already does, introduced to each other.
Here’s the problem it solves, because it’s one you’ve definitely had.
You’re working on a site. It’s not ready. Visitors are landing on it anyway — seeing unfinished pages, placeholder text, a theme that isn’t quite right yet. You need a friendly “we’re getting there” page in front of everything while you keep building behind it. Simple.
The existing solutions are, almost without exception, overkill. Branding splashes. Settings panels. Countdown timers. Asset bloat. Fonts that don’t match your theme. A footer credit pointing back to someone else’s website on your coming-soon page.
I didn’t want any of that. I wanted something that looked like the site it was protecting.
So I built one. Sort of.
I should be honest about what “I built” means here. I had the idea. I understood the problem. I knew, roughly, that WordPress has a way to intercept a visitor’s request before the page loads — and that if you catch it at the right moment, you can redirect everyone who isn’t logged in without touching the theme. I just didn’t know exactly how.
That’s where AI came in. Not to do the thinking. To explain things.
Think of it like having someone next to you who has read every piece of WordPress documentation ever written and will answer your questions without making you feel stupid for asking. You still have to ask the right questions. You still have to recognize when the answer is correct. But you’re not starting from zero every time you hit a wall.
— Claude (Sonnet 4.6)
There were things I genuinely didn’t understand — how WordPress handles the redirect hook, how to create a page programmatically and clean up after yourself, why you escape output even when you wrote it yourself. I’d ask. I’d get an explanation. Then when the reviewer came back with notes and guidelines, I’d forward the feedback — and this time, the conversation was different. Not because the AI remembered anything. Because I did.
That’s the loop. The memory doesn’t live in the LLM. It lives in you. The AI explains; the reviewer teaches; you own the result.
And the most difficult part of the whole process? Finding a name that the plugin directory would accept. That took longer than the code.
The plugin itself ended up beautifully small. So small, in fact, that it doesn’t need modern image formats. No AVIF, no WebP — just an ugly, completely customizable gradient that somehow looks intentional. That’s the whole aesthetic.
Activate it, and it creates a WordPress page. Visitors get redirected there. Logged-in users see the normal site. Deactivate it, and the page becomes a draft — saved, not deleted, ready for next time. Nothing left behind.
The page uses native blocks. Your theme’s fonts. Your theme’s colors. You edit it the same way you’d edit any other page, because it is any other page. The social links block, if you leave the URL fields empty, renders nothing at all — so the default page works without customization, but rewards you if you take five minutes to personalize it.
No settings panel. No configuration. No maintenance burden you’re signing up for without realizing it.
I want to say something about restraint, because I almost blew it.
At one point, I started exploring the Gutenberg Data Views API — @wordpress/dataviews — thinking I could build a nicer way to select templates from inside the plugin. Grid views, variants, the whole thing. It was genuinely interesting. It was also unnecessary for a plugin whose value proposition is that it requires no user decisions.
Even variants were too much. So I stopped.
This matters more than it might seem. There’s a version of this plugin I could build that would take weeks and require ongoing maintenance and slowly drift away from what makes it useful. The discipline isn’t in the building. It’s in the not-building. Don’t add something you can’t explain. Don’t add something you can’t maintain. If it starts to feel like overkill, it probably is.
Core needs connectors, not competitors.
The plugin Almost Ready Temporary Page passed review, and it’s tested up to WordPress 7.0. The graphics were made with AI, which feels consistent with how everything else came together: human idea, machine assistance, something that actually ships.
If you’ve had an idea sitting in your head for two years because you assumed you’d need to hire someone to build it — this is your sign to reconsider.
— Jos… and Claude.
Leave a Reply