As I’ve shared before, I’ve been building software called PageMotor that is going to replace WordPress—and perhaps many other established “solutions”—in popular usage.
The basic premise is that nothing has been built to leverage AI in the nearly infinite ways it can be leveraged, so that’s what PageMotor is all about. AI as a fundamental architectural consideration.
But that’s all boring stuff. The real point is doing amazing things that were not possible before… in a fraction of the time.
PageMotor has now redefined the traditional agency workflow:
Idea (AI-generated) design live, working website
… by condensing that days- or weeks-long process down to a few minutes
It’s remarkable enough that AI can now generate extremely competent web designs. (And if you’re into this kinda thing, Claude is the reigning champion here.) But the challenge lies in what do you actually do with that design once you’ve got it?
PageMotor is the first platform to enable you to use these designs in an efficient way; it actually brings them to life!
Check out the video above to see how a historically technical process is now trivial. (In fact, it can now be run by AI agents and doesn’t even require human intervention, but that’s a conversation for another day…)
It’s all going way above my noggin but I’ve worked with several web developers who would love any advantage they can muster against the sweat shops out there.
Yesterday I went on a picnic with a few families, and I happened to talk about this topic with a German web engineer.
We all agreed that in the AI era, the way information is accessed is changing. Traditional companies’ reliance on social media exposure will gradually shift toward placing greater importance on well-structured websites. This has a lot to do with the success of Gemini and the growing dependence on AI in general. While AI can search social media content, it still has limitations. That’s why we believe websites will regain their importance and be revalued by businesses of all kinds in this new era.
However, in terms of visual aesthetics and truly human-centered UI, I believe designers will still be essential in the short term. I’ve signed up for the email list and am really looking forward to how PageMotor develops.
I’ve long been a WordPress enthusiast, but I often run into frustrations when integrating plugins. While the vast ecosystem of plugins is one of WordPress’s greatest strengths, the sheer number of compatibility conflicts is also its Achilles’ heel. Another constant headache is figuring out how to properly adapt a site across different screen sizes (phone, tablet, PC).
This is true! Optimizing content to be cited as “answers” by AI is a huge deal moving forward. It’s going to mimic the SEO gold rush of 2005–2010… but this time, it’s going to be a lot more useful (and a lot less gamey).
Nice! I’m glad you’re coming along for the ride.
There’s so much here, but suffice it to say, “patternized integration” is a big part of PageMotor’s architecture. In fact, PageMotor’s Plugin model is the “secret sauce” that elevates it above everything else.
When you create a WordPress Plugin, you get nothing. WordPress doesn’t extend any native functionality to Plugins at all; it’s up to the developer to decide how to integrate with the Platform.
Freedom is great, but patterns are better. PageMotor knows there is an array of key actions Plugins will want to take. To accommodate this, it sets up something called valet methods that enable Plugins to provide a bit of (highly patternized) data; PageMotor then takes this data and performs expected actions in response.
Take an options page, for example. Within a WordPress Plugin, a developer would have to:
Use something called hooks and filters to deliver an options page in the Admin at a chosen URL
Deliver ALL the HTML for the options page (using whatever markup conventions the author chooses)
Deliver CSS to make the options page display as desired
Use hooks and filters to save any options data entered on the custom page
Use hooks and filters to retrieve this data when and where appropriate
Now contrast that with PageMotor’s Plugin valet methods. Let’s say you wanted to add an options page in the Admin:
In your Plugin, specify an options() method and have it return an array of options in PageMotor Options Array Format
That’s it. The options will appear in a pre-specified location in the PageMotor Admin; the page will render as expected; all options will be saved/retrieved automatically—and precisely at the right time—with no additional work from the developer.
And this is just one example of the ways PageMotor takes the chaos of WordPress and finesses it into a brilliant orchestration.
The current Beta participants all agree, it’s pretty much impossible to return to WordPress after playing with PageMotor. Although the platform is multifaceted (like WordPress), every facet you interface with is, comparatively, a delight. (And the speed alone makes it nearly impossible to tolerate the clunky slowness of WordPress.)
I’m not sure if my understanding is accurate, but I see PageMotor as a new-generation CMS that uses AI to rapidly extend and assemble the functionality a site needs. somewhat like Vercel combined with a headless CMS？
This leads me to another important question: when a website needs to integrate with banking or payment systems, can PageMotor offer the same level of automation while still maintaining strong security once AI is involved? In that sense, is PageMotor’s admin backend designed as a sandboxed environment, or is it—at least to some degree—a centralized, enterprise-managed platform?
Once a website is connected to a banking or payment system, the discussion shifts from a feature checklist to a question of architectural trust. Is this a system that I can reasonably trust with real commercial responsibility?
I’m preparing to update my portfolio and hope to launch it sometime this summer. This approach has made me seriously reconsider whether I should completely move away from WordPress and try something entirely new.
Exactly right. Only real distinction here is PageMotor runs on every minimally-equipped Apache server on Earth, and everything is in one location (MySQL or MariaDB runs on the same server). No “cloud connections” or other stuff required, even if behind the scenes.
The easiest way to think about PageMotor is to see it as WordPress… except it doesn’t suck, is infinitely more flexible, and is designed around JSON objects as the primary data structure (which creates natural fluency with AI).
With ecommerce, the standard is for websites NOT to collect any payment data or personally identifiable information beyond emails and, if provided, names.
I’ve encountered a few government systems over the years that had a significant WordPress architecture, and in almost every case, they used a completely custom wrapper around the portions of the system that handled sensitive government or banking data.
Outside of that, the aforementioned standard is the norm. Individual payment data (credit cards and the like) lives with Stripe, PayPal, etc. and is never stored with the site in question.
At a glance, I don’t see any major reason why PageMotor would ever be excluded from this type of application, but there are inherent risks whenever a platform is using a typical DB connection and serving content directly from the DB (with some processing, but still…).
Interestingly, I’ll be able to use OpenClaw to try and breach PageMotor while also offering recommended hardening techniques, so that will be a fun side quest once the platform is more mature.
But the “big idea” is that PageMotor will be a natural architectural foundation for any application that requires front end presentation and an admin to control things.
I think if you were to try PageMotor, you would be very grateful you did.
The AI-native aspect is producing a whirlpool of increasing velocity where PageMotor can achieve more in less time.
Three weeks ago, I went from an AI design to a client deliverable in 6 hours. This week, it’s down to about 30 minutes.
Right now, I’m drafting documentation the AIs will use to create Plugins for the PageMotor ecosystem. I suspect that within a couple weeks, I’ll have a wildly convincing prototype of a forthcoming ecommerce Plugin that will replicate months of development work in seconds.
After that, I’ve got a plan to teach the AIs how to remix all of PageMotor to serve as the core piece of software for an unbelievable array of industries.
Not many people realize this, but most of the “software business” is bespoke software + sales teams that get deployed to dominate specific verticals.
For example, I own a yoga studio. The studio pays $300 per month to this one company for their yoga-specific website + studio booking software. That company is worth over $30M.
But the software is objectively terrible. It’s also simple enough that it doesn’t warrant its own custom build-out for this specific industry. After all, how much different is a yoga studio from, say, a barbershop? Software-wise, the difference is quite small.
PageMotor has been built in such a way that it can be remixed and white-labeled for industries like this (and well beyond—think law firms, doctors’ offices, etc).
I’m going to teach the AIs how to do this, and then it’s off to the races to take over verticals with generationally superior software.
It’s not publicly available yet, but I’ll be re-opening the Beta again soon (within the next 2 weeks).
Since your sister works with WordPress, it makes sense for her to follow along with the PageMotor email list to keep up with what’s going on. She’ll also get Practical AI updates this way, which is probably where the video you saw was from.
My perspective has always been rooted in the design side of building things, so I tend to approach problems from a non-programmer’s point of view. In the past, that sometimes caused frustration for developers. Even though the process could be demanding and time-consuming, we always managed to deliver the work successfully in the end.
If there’s a tool that can lower the barrier to coding while also offering a more human-centered interface, I think that would be a truly killer application.
Thank you for your detailed explanation. I’m really looking forward to the day the system is completed.
Out of curiosity, which AI API is PageMotor built on?
The Agent connections in PageMotor are open-ended and can accept connections to any AI with an API. (But the stock connection for now is Claude—just enter an API key, and you’re good to go.)
There’s a 2-pronged approach here that catches people off guard:
Some AI functionality can happen in the PageMotor Admin (Architect AI) and on the front end of the website (Accelerate AI). This requires the API connection described above.
Other functionality, such as the production of a design and the generation of an import file to make the design appear on a PageMotor site (as shown in the video above), must still happen on the AIs themselves (i.e. within Claude, Gemini, etc.)
I guess I’m just taking an adversarial attitude towards this by default because I’ve adopted the anti-AI sentiment held by many people right now, but this honestly doesn’t seem so useful If this is about lowering the barrier of entry to a web presence, that’s already been solved with platforms like Squarespace and Shopify and whatnot. And if it’s about a zero-effort CMS… well, what would this add over wordpress? All it’s doing is closely integrating the work of some LLMs. I feel like you could achieve as much with GH copilot.
So I’m doubtful, but tbh my experience doesn’t go much beyond static site generators and such. I might be misunderstanding the use case or something. Also did you write your initial post with AI? lol
The current agency design-to-delivery process takes weeks. What I’ve demonstrated here takes minutes. That’s a business-changing difference.
Agencies also typically deliver websites on WordPress, Shopify, SquareSpace, or WebFlow. Each of those platforms has glaring negatives that include pricing, complexity, extensibility, performance, SEO, and beyond.
The process I’ve shown here works with a new platform called PageMotor that eliminates most (but in fairness, not all) of the problems associated with those other platforms.
I understand the sentiment here, but it looks ridiculous once examined in detail. Clients need websites they can manage. In practice, that basically means one of the platforms I listed above.
(In most cases, real people have to run and update websites after the hand-off from the agency.)
PageMotor is a direct descendant of the CMS lineage of the past 20 years. It has been engineered to address every touchpoint on the spectrum, from demanding developers to normie end users. Instead of requiring a complicated technical setup with a litany of dependencies, it runs on the most basic and ubiquitous hardware connected to the internet (any minimally equipped Apache server).
And its core data model relies on JSON, which is the “data language” LLMs speak natively.
Bottom line: PageMotor is positioned for the future and can provide leverage that no other platform can match. The demonstration of a weeks-to-minutes efficiency improvement in the typical agency workflow is one example of this leverage, but it is by no means the full story (which will continue to unfold throughout 2026).
One of my companies, Verdon Partners, could possibly use something like this. Your explanation of the agency model and build process is something that I’ve completely collapsed into one os. Please email me access once it’s ready to try.
(AI-generated) design 
If this is about lowering the barrier of entry to a web presence, that’s already been solved with platforms like Squarespace and Shopify and whatnot. And if it’s about a zero-effort CMS… well, what would this add over wordpress? All it’s doing is closely integrating the work of some LLMs. I feel like you could achieve as much with GH copilot.