Show HN: An AI-Powered WordPress Site Builder That We Are Open-Sourcing Today
themeisle.com
themeisle.com
For others, it's "Full Site Editing" themes, which means... I guess you can edit anything in the theme customizer?
That's a very good question. In WordPress, Full Site Editing (FSE) allows you to design your entire site—including the header, footer, and everything in between—using blocks. Blocks are the components for adding content in the new WordPress block editor, enabling you to easily incorporate text, images, videos, and more.
For this new approach, your site will require a Block theme, designed specifically for compatibility with this new editor. Also, one thing to note here is that the traditional theme customizer is no longer present in this environment. Instead, FSE allows direct manipulation of site elements within the WordPress Block Editor interface.
There are some high-quality GPL-spirited plugins and themes (like those by Anders Norén https://andersnoren.se/teman/ or Oliver Juhas https://www.webmandesign.eu) amid an utter avalanche of SEO cruft so you have to know where to look.
These are us programmers' trappings. The users don't give two sh*ts about that. They care whether what they use is easy and it works. And Wordpress is both of them.
None of the 'better coded' competitors have been able to eat into its market share ever - demonstrates the gigantic gap between what programmers think to be important and what users think to be important.
> An earlier post here on hn by dan about web bloat seemed to show wordpress was actually the best option in that category?
Yep, that's the irony. All the competitors kept adopting 'better coding paradigms' as their respective fads hit their maximum, ending up in bloated stuff. They occasionally had to break backward compatibility to get rid of the code that was adopted in that fashion too - doing much harm to their ecosystems.
But WP went for ease of use, simplicity, and from another angle, having to protect backward compatibility that WP project gave so much importance helped it keep itself lean, mean and simple. So now WP is something you can just launch in a total of 2 second from zero to production-ready thanks to one-click installs, hosts and all the infra providers that offer it, with ridiculously low initial and ongoing costs. Also because it didn't buy into the recent frontend fads and stayed server-side rendered, it still remains very lightweight and renders in an average of ~1.5 seconds in any device/browser. Of course, some of the framework craze made its way to wp in the form of React and full site editing, but still, the general landscape remains lightweight especially if you use hybrid themes. (they allow all the fse and page builder functionalities by trading off only not being able to edit header & footer by dragging & dropping, so its a good trade-off).
In a way, WP having to deal directly with the actual users whose priorities were always on the practical side helped it a LOT in getting the market share it has and the reliability, extensibility and ease of use it still excels in.
Once upon a time, I was involved in creating the largest WordPress media platform in the world with WordPress. You can view it at news.com.au, but there are about 20 other publications hosted on the same platform.
The process goes like this:
1. Login to update plugins.
2. Some plugins don't auto update because they require licenses.
3. Search for the theme's proprietary update process or plugin.
4. Discover that the Envato support license is expired.
5. Renew license (typically 6 months).
6. Update plugins again and... pray that it works because support is always a gamble.
We stay away from Envato and the like if we can, but inevitably we acquire clients with themes purchased this way.
With all the awesome open source self hosted apps these days I'm surprised there's nothing better for generic websites.
Ended up using only a single third party plugin (Polylang), and doing the other stuff we need in a custom theme based on Automattic's (possibly a bit unmaintained) _s base theme.
Feels like a good decision, only took me about a day more to fiddle with the CSS to make things look right and write a bit of PHP code for some functional requirements we had. Seems way more economical than dealing with a soup of plugins and some shady theme. Since you can nowadays just create custom blocks for all the site elements, it's a pretty OK development experience.
But I also can't quite believe this is the state of the art... I use static site generators whenever I can, but for a bit more dynamic websites, it seems WordPress is one of the best out of what seems like only bad options.
Also an interesting case: Automattic seems like a company that's set up to resist enshittification. But all those theme and plugin authors with profit motives are not, which has about the same effect as if Automattic was ruining the platform for short term profits.
What??!? Automattic took VC money. Not those small time plugin and theme authors who are trying to make a living by doing open source. When those smaller plugin and theme authors want to exit the business, they sell their businesses to people who are inside the WP space instead of investment funds or VCs. They keep things inside the ecosystem.
Whereas the 'not enshittifying' Automattic is trying to get money out of its entire ecosystem by doing a lot of shady things - from flooding google with plugin listing links that go to wordpress.com instead of wordpress.org so that those who want to use plugins will have to buy the $300/year Automattic hosting plan instead of being able to download the plugin from wordpress.org and use it at a $5 dollar hosting that they could get from anywhere. All the Automattic plugins are full of popups for upsells and upsold features too.
Its amazing how people have a lot of grand-standing opinions without accompanying levels of information.
Its much better to have that 'hive of scum and freemium' and pay some money to those plugins than having to look for another plugin to build your site on every few years.
I’m asking because for this use-case Wordpress Multisite seems to be the only option, but I’ve also always despised its structure and inner-workings.
However, for 300 sites, anything but a custom web app will get out of hand because that’s too many to manage in a UI that’s primarily meant for a handful of sites.
We've also been looking at a multitenancy option for WordPress, in particular using Gridpave https://gridpane.com/kb/using-gridpane-multitenancy/ but you need the $5000 annual dev pro plan (annual for unlimited sites). We use Gridpane already on their lower cost plan and have found it excellent value for money.
We will probably go with Payload. WP is getting messier by the day. With the AI coding tools getting better and better it feels like having as much flexibility as possible will have a huge benefit down the line.
Last month I created a brand new site using WordPress. I followed a 90 minute video on YouTube where I was instructed step-by-step on what to do. The result is that I'm happy with the finished product and learned a lot about WordPress along the way.
However, I would have NEVER got there on my own absent this remarkable YT video. WordPress is too complex, too bloated, too filled with messages from plugins, too reliant on magic, too reliant on plugins to do anything advanced beyond setup a post or page.
I hate WordPress so much. I'll still use it because it's entrenched, and because it's less time that coding a Jekyll site to equivalent functionality.
But I hate it. Thoroughly.
I wouldn't count on that. It required a 90-minute video for you to get going with Wordpress. I would say, if you know some basic stuff (e.g., know how to install gem, how to write markdown/html), you can set up Jekyll in under 90 minutes:
gem install jekyll bundler; jekyll new myblog; cd myblog; bundle exec jekyll serve
Now, you have a server up and running. You can start creating a post by creating a markdown file in the _posts folder.
For me personally, it would be faster to learn any easy-to-use static site generator than Wordpress.
For a lot of people clicking pretty buttons is easier to understand.
Cause it seems like building any site with a tool you don't understand and don't particularly want to understand is going to be a bad experience?
(I see, my original comment was downvoted -- I've no idea. We started WordPress with no option to have multiple pages and now see where we are and what we can do with it. Isn't that an interesting titbit.) :-)
I had this mindset for years. "It's not real programming. It's for people with no brain", I would tell myself. A lot of other developers share this idea, I think.
However, I currently am in the plan of launching a series of blogs, somewhat related to each other. I figured I'd try out wordpress, just to see what it's like, and also because I plan on having other people working on these blogs.
Whoah! Completely blown away by wordpress and how quickly I can set things up. PHP is quite simple, and the built-in functions are very helpful.
If you plan on just launching some products, learn wordpress. If you want to learn a programming language, go ahead and write the entirety of the code base yourself - that's the only reason I can think of to not use wordpress (for myself).
My favorite thing was how easy it was to create a simple plugin from where I could inject my own php / html / js/ css at just about any stage of the pipeline, and crazily enough, I tried too many cms that would not allow me to override and direct edit the html on a page. It does win in the "easy to deploy but fully customizable" category, just because of that.
But my thought when using it was sometimes "they made this more complicated and difficult than doing it directly, because they managed to convince people the direct programming was out of reach". But that is how I feel about most frameworks, so I might not be objective on that one. Overall it was faster, but the tradeoff was massive opacity. I recently noticed a tracking pixel on my site and have to figure out which a**hole plugin dev is injecting it, then find a different plugin or write my own. I have always been vanilla for my other sites, so learning to combat other vendors and having an adversarial relationship with my own site really breaks the mental model for me.
1. You can install it in a toaster. The cheapest hosting service will offer PHP and MySQL. It will run slow, but it will run. It will be easy to install and start working right away.
2. You can train anyone to use it and put content on it in an hour.
The security industry in WP space is bigger than most open source spaces, comparable with Linux. If you go use plugins with a few hundred installs, yeah, you may end up with security risks. But if you use plugins with a few thousand installs and more or use plugins that are from maintainers who have other plugins with such large amount of installs, you should be safe.
As an entrepreneur I see why WordPress has such a large command. Plenty of people who can work on it. If you are nontechnical and use a hosted solution you can create a good enough MVP with all of these plugins. As an engineer everything about it is annoying but it commands 42% of the web so it must be doing something right.
Yeah. I also had the same mentality until I started interfacing with actual users, site owners and customers and they started telling me that they don't give two sh*ts about the things that we programmers think to be 'important'. They care whether something is easy, cheap, it works and that's it. And they are right - they have actual businesses to run, and their livelihoods depend on it. Then I tried setting up and using a few WP sites - holy crap, that was easy. Even if the coding paradigms in some parts of its code were not 'up to date', it worked like a clock, and it took away all stuff that I had to do manually before. Then with time I understood that the project was not changing those 'not up to date' code inside the codebase to keep backward compatibility - which is critically important for the users (who now included me), and I came to appreciate it. I also did some coding on it and got some users and learned that users really, really, REALLY hate their websites and businesses getting disrupted because some programmer/maintainer thought to update something in a backward compatible way because it was 'better programming'. So now I don't do it either.
With me it's the exact opposite. Jekyll, a GH actions, deploy to a static site hosting takes under an hour. I know Jekyll, so it's a bit cheating, but yesterday I did the same with hugo, that I never used before and within 1:30 I had a site running on Digital Ocean.
And It's also not because I don't know WP. I've build and scaled a dedicated WP hosting company so I know quite a bit about setting it up :)
In my case it's always because I know the primitives of the web very well: HTML, CSS, JavaScript (and the JS/HTML Web APIs).
WP has layers upon layers upon layers between me and these primitives. Hugo or Jekyll don't: they are as close to these primitives as possible. I know the primitives, I never know these layers upon layers upon layers. They are shifting targets, poorly documented and often horribly executed. So instead of "just using flexbox for this thingy", In WP I need to learn yet another "site builder", or "theme framework", or, worse but rather common: all combinations of all options; infinite amount of permutations that interact, or conflict, or both.
(And yes, I know I can build a WP theme from scratch, staying close to the primitives. But that's far, far more work than 1:30h)
This is just the startup cost. Which is, in my experience, rather high with WP. The real cost comes at maintaining and future development and scaling.
Like I said: I built and scaled such a hosting.
But while you can start in just minutes, you don't have that publishable site that you have in your mind, in minutes. It takes hours to learn where to find what. To select a good theme. To find the right plugins. To remove the wrong plugins. To then fix some error that comes from removing the plugin. Or to fix an error from installing the wrong one. Experienced WP devs even ask hours for this: people who do this all day, for a living, will charge you hours to build this for you.
Sure, you may be lucky and have some goal in mind that happens to be ridiculous easy with WP. Or you may not have something in mind and just go-with-the-flow to end somewhere that happens to be easy with WP (a good strategy really).
But, in general, I have something in mind. A landing page for a startup. Or the outlines of a webshop, or a simple blog even, and it will take me hours or days banging against WP to get there, whereas with hugo or jekyll that's less than a few hours of banging.
Now try doing that as a non technical person with anything else and report to us if you succeeded. You won't. This is why WordPress is popular. Liking or hating doesn't matter. No one has been able to build something that can replace WordPress.
I warn people that even though a good and easy to use tutorial like XYorZ may help teach, be wary that if they are teaching using elementor, bricks or similar page builder (added plugins), that it may be good knowledge, but that's not really wordpress per se.
Anyhow, with the new FSE editing options, I will say that I finally feel we can put the separate page builders to pasture for the most part, and all the bloat that comes with them for most use cases.
Super Appreciate the vids that Jamie Marsland has been releasing - many showing how to do nice designs with base WP, watching the process is helpful getting up to speed with current options - and shows how modern stuff can be crafted without tons of fancy addons these days.
Navigating the new FSE stuff is very different than previous WP ways like the customizer or 'additional css' blocks, it's finally getting noob friendly again.
If you needed a 90-minute video — I daresay you either did it wrong, bit off too much to chew, or were better off using a high-code framework instead.
One of the reasons WordPress got popular is its at-the-time famed ‘five-minute install’. [1] You just throw the files on a server with PHP, give it a MySQL database and that’s it. (Nowadays, even this five-minute process is usually abstracted away behind a mouse click.)
With that, you get the core WP functionality, i.e. a modern theme, the ability to create blog posts, pages, and have visitors leave comments. For many people who just want to get a relatively simple website, this is quite enough, and IMO there is still no comparable no-code site engine.
Want a different look? Just search on the official theme directory (quality varies, but there are decent ones — just try not to get forced into the old ‘classic’ editor), use one of the first-party themes, or write your own.
Want to extend the core functionality? Just search on the official plugin directory (again, quality varies, so exercise judgment) or write your own.
Now, is it possible to blindly throw a cocktail of plugins at the wall, resulting in an unmaintainable abomination? Yes, it is — but as you’ve already discovered, you won’t necessarily like the result.
[1] https://developer.wordpress.org/advanced-administration/befo...
Huh?
The out of the box Wordpress is everything one needs to publish things on the internet. Its easy as hell and that's why its 43% of the websites on the planet.
If you BUILT a site, then it means that you did use it for a specific purpose other than just blogging or having a presence on the internet. If that is the case, how did you expect it to happen? Without zero effort on your side? Have you tried doing the same in React or any modern frontend for comparison?
Existing AI tools alongside other open-source builders (Puck for React, Primo for Svelte, etc.) equals all the good things about WordPress's "WYSIWYG-ness" minus all the unnecessary bloat that's piled up over the years.
[1] https://kinsta.com/wordpress-market-share/ or https://trends.builtwith.com/cms