This week we’re demoing our first proper test: Can AI build a real page for a design studio inside of Elements; a hero banner, services section, contact form, the lot, all from a single prompt? Watch the video to see how it went.
As always, these weekly dev diaries are an open look at the research and development behind AI in Elements. Nothing here is shipping yet, we’re sharing the work as we go so we can get feedback and you can see what we’re actually working on.
Best of all, this is just the beginning, we’ll be back next week with another AI update.
We’d love to know what you make of it so far. Does this look genuinely useful, or are we missing something? All feedback is really appreciated
Quite impressive, especially for new users of Elements and could navigate them around some of the pitfalls like setting up a CMS and making sure the page is a php, a common newbie overlook.
I think the other thing is the Ai will probably know what all the settings in the component inspector actually do. Rather than my have to try them to figure out what they do.
I like the previews and you would be foolish not to address this. Ai is not going away and it will overtake those that aren’t preparing just like the video rental stores who didn’t prepare for the outbreak of streaming media.
Once this AI-assisted design is released, I’ll finally start using Elements to redesign my websites but right now, I don’t have time to manually recreate them going from Rapidweaver 8 to Elements, so very much looking forward to having this AI feature.
The video strikes me as impressive because, at its core, it’s handing over to Claude AI the ability to operate the main application — in this case, Elements — to perform tasks that previously required manual intervention from you. This allows you to focus on fine-tuning exactly what you want to change and infusing your own creativity into the final result. That’s certainly an advantage… but also a risk. I hope my perspective isn’t interpreted negatively; it’s simply grounded in my current experience — which is virtually identical to what’s shown in the video. I stopped using Adobe CC and switched to Affinity by Canva. This tool suits me far better than the previous one, as its features align more closely with the kind of work I do. Recently, they introduced precisely what’s demonstrated in the video — but using Affinity: you’d tell Claude AI what you wanted to do in your design, and Claude AI would execute it automatically. It was truly remarkable. Yet, within the first week of its launch, the community began rapidly sharing numerous prompts and ideas — right? — that users could exchange, eliminating the need for each person to build them from scratch. That practice quickly snowballed, and within three weeks, a publicly accessible gallery had emerged, packed with dozens of ready-to-run prompts and scripts for a wide variety of tasks. As a result, Affinity effectively transformed into a secondary tool: Claude AI became the primary engine, while Affinity became more of an interface layer dependent on it — without Claude AI, many of the most impressive features available today simply wouldn’t exist.
What I want to emphasize is that Elements was created precisely to achieve autonomy: its very purpose was to stop being a secondary tool and instead reclaim its role as the primary application — not a mere shell — for building web pages. That earlier dynamic was counterproductive, and that’s why Elements exists today: with the explicit intention of reasserting itself as the central, indispensable tool. However, opening the door for Claude AI to assume core functions risks repeating the same pattern — giving rise to yet another community — as already happened with Affinity — where users share prompts, commands, and scripts to tailor the tool to their specific needs, but once again pushing Elements into the background.
In summary: No one asked me for advice — but I’m offering it anyway. Let’s harness the power of AI as a resource we don’t access directly, and keep Elements firmly as the central tool we actively use to access those benefits — whether via Claude AI or any other available model. That way, Elements retains control and remains the primary application, avoiding the repeat of what happened with RapidWeaver and third-party elements. This approach could even work better than your own current workflow: you could simply ask Elements directly to perform a task for you, rather than opening Claude AI, working there, and then seeing the result reflected back in Elements. I hope this viewpoint comes across clearly — I recognize it may be highly polarizing.
This is amazing, Dan. Maybe this could be the way Elements imports Rapidweaver sites with a better result? Or maybe we can paste the code for a site to the AI and ask it to build it using Elements. A time and effort saver to be sure.
And, frankly, kudos to you and the team for being able to hook so deeply into Elements so that the AI LLM understands what Elements has and can do. Amazing stuff. Keep it up!
I love what I’m seeing here. Adding this capability is going to be a game-changer for Elements. I’m a member of several SEO communities, and lately I’ve noticed a growing push toward static-based sites. People are tired of the bloat that comes with WordPress. A lot of them are already experimenting with AI and Astro to build sites, and the results look great, but as @dan pointed out, they end up completely dependent on AI just to make changes.
What I see happening with Elements is the best of both worlds. You can build with AI if you want to, but you still have full control of your site without needing AI involved at all. Once this gets implemented, I’ll be recommending Elements as a web-building solution across all my groups.
That is going to be interesting for us Pros, when our customer could now point claude to an existing website and tell elements to recreate it - or build a new one with new design but take over the contents.
But I really like to be in control of the overall project and intervene and change as much as I like.
This is a valid observation. I see the line. Dan mentioned something that I think is critical to remember. AI building a sight does not take ownership away from you. It’s like hiring a web designer and coder amped up on Red Bull. When AI is done you still own the doc and you can still edit it.
Yep there will likely be an AI sub market. And that sub market is more than likely going to make third party developers questions the validity of putting in time and effort to make a product. However, AI is not taking ownership of Elements App from Realmac nor is it taking the project away from the webmaster.
Truly amazing. I can see why you’re excited about it. As you like to say Ben - “Very powerful”
It looks like the future is disparate tools/apps manipulating design, graphics, CSS/HTML etc. all crafted by AI with the web developer being the captain of the ship who sets up the master course and then finesses things as the project evolves.
“Change all headers to serif fonts” – “Make the logo 20% bigger and change the color palette to a Winter vibe.” - “Make a Spring themed version and replace current site on April 1.”
“Add ‘special-dinner’ banner template. Event starts May 3. Remove this banner June 1. All content/design hints in ‘May-din-din.pdf’. Send event info emails & texts to VIP-list once active. Encourage use of ‘buy tickets’ dialog”
Maybe you can use AI to write Help Pop-ups for all the Elements tools and features that can be automagically updated when you change interface or when you change how things work. In the help Pop-up put Link to detailed information.
And, of course, a dialog where we enter our credit card info so that we can pay for the thousands of tokens we’ll burn thru because it’s so much fun and time-saving. In fact you could lie in bed, sucking on bon-bons, looking at edits on a ceiling-mounted 100" screen. “Publish site when you detect I am asleep.”
Hi Dan. I really appreciated this video and am very exited about where this could lead. First, I’ll give you a brief background about me, so you have some context for my use case. I am not a web developer, but I have experience building a very small number of websites for my own businesses or close partners. I fell in love with Rapidweaver more than decades ago after trying to use software like the old Dreamweaver or relying on third-party services for development needs. Rapidweaver allowed me to build and maintain good websites without having to get special web development training. However, the one thing I always struggled with is how to get started with a new site and determining what type of content I could, or even should, include. Tools like templates and Stacks certainly helped, but it was still always difficult for me, being someone that doesn’t build sites for a living.
I recently purchased Elements and have done some initial work to familiarize myself with the software, but have yet to build a site with it…partly because, even though Elements (and Rapidweaver Classic) make it easy to build a maintain a site, it still feels like an overwhelming task getting started. If I were able to use Claude to help me get started and to help me with some of the less desirable aspects of site development like SEO, I would absolutely love it. I have two existing sites originally built with Rapidweaver Classic that I want to both convert to Elements and upgrade the content. Based on your video, I am fully convinced that Claude could accomplish all of this in an incredibly short amount of time and that make me very excited about what you are building. I have been using Claude the last 6-7 months to build apps as pet projects and for job-related tasks and have been blown away by what I’ve been able to accomplish.
Of course, the power of AI has made me consider relying on it instead of something like Rapidweaver Classic or Elements to handle my small website needs. However, as noted in your video, the process and results are clumsy to work with. While AI can be very helpful with designing sites and content, as you also mentioned in the video, I still need full control over the site…and I would add, I also need the magnificent simplicity Elements provides for actually controlling that content.
Unfortunately, this is an optimistic forecast; if you raise the bar a little, the time will be reduced. Then we also need to see how much the extra work outside the plan will cost in this case.
Absolute rubbish! Just like Tesla FSD for Elements! Also @dan you need to work on you “moaning voice” about people like me complaining about you not doing other things - didn’t sound like me at all!
Ok so I am now off in my Tesla on full FSD so I can watch some more of these incredible Dev Diaries you keep putting out. I’ll sit in the back seat with Claude and let him finish in time for to watch the footy tonight! I can only see the good in this - more time to watch footy whilst getting work done! RESULT - COE!
Just wanted to chime in and let you know we’re reading all your feedback and discussing it internally as we progress with implementing AI into Elements!
We will of course be back next week with another AI dev video
I used Linguix AI to fix grammar and spelling
this was possible now…
absolutely agree - you need to get this up and running quickly or get left behind as you stated…
