Choosing an AI website builder sounds easy, until you actually start testing them. Almost every tool promises to turn a simple prompt into a polish...
For further actions, you may consider blocking this person and/or reporting abuse
The split you draw between "generates a page" and "generates a working product" is the most useful framing in the post, and I'd put it even earlier in the decision: most people compare builders on the first draft, but the cost shows up later when you need to change something specific.
Two things I'd add from testing these tools:
Curious whether you tested the form and commerce features on the free tiers, or only on paid plans? In my experience a few of these gate forms and custom domains behind payment in ways that change the comparison.
Thanks, this is a fair point on exit paths. I should have made that a separate criterion in the post. Lock-in was something I noticed mostly after the first project got bigger, so it's useful to see it framed up front. I agree that the export question should come before the design question. I'll consider adding it as a sixth test in a future update.
On the second edit test: I did run follow-up changes, but I didn't document the results as carefully as I should have. The tools that held up best were the ones where a second change didn't undo the first. That's a good test to include.
On the free tiers: I tested forms and commerce on the free plans where they were available. Some features, like custom domains and certain form submissions, were gated, and I noted that in the pricing sections. You're right that gating changes the comparison, so I'll make that more explicit.
One gap I'd flag: the post treats "AI website builder" as one category, but the risk profile differs a lot by what the AI is generating. Tools that output editable HTML/CSS or React give you something you can review in a diff, hand to a developer, or host elsewhere. Tools that generate into a proprietary visual editor make it harder to see exactly what changed between iterations.
That matters for SEO and accessibility, which the FAQ mentions briefly. AI-generated markup often ships with missing alt text, skipped heading levels, and generic titles, and those are much easier to catch and fix when you can inspect the source. I'd run a quick Lighthouse or axe check on the first output from any builder before you spend time refining it.
Also worth noting for anyone on a budget: pricing for "starting paid" plans often excludes the custom domain, email, and the credit overages that kick in once you iterate for real. If you're comparing costs, model a realistic month of editing, not just the sticker price.
Good catch on the category problem. "AI website builder" covers tools with very different outputs, and I grouped them more by use case than by how the output is stored. Tools that produce readable HTML, CSS, or React are much easier to audit, which affects SEO and accessibility work.
I like the suggestion to run Lighthouse or axe on the first output. I didn't include accessibility testing in my process, and that's a real gap. I'll add it.
On pricing, the point about credit overages is well taken. The sticker price understates the real cost for anyone who iterates heavily. Modeling a realistic month of editing is a better way to compare, and I should have said so in the post.
The line that hit me: "AI can dramatically reduce the time needed to create the first version, but it does not remove the need to review, edit, test, and maintain the website."
I'm a beginner — three weeks into Python, writing tutorials on Dev.to. I don't have a business website or a SaaS product. But I've been watching the AI website builder space closely, and this comparison is one of the clearest I've read.
What I appreciate most is that you tested them against the same five questions. Most "best AI website builder" lists just show screenshots of pretty generated pages and call it a day. You asked: what happens after the first draft?
That's the question that actually matters. I've tried a couple of these tools for fun. The first generation always looks impressive. It's the second version — when you actually need to change something — where the cracks show.
The grouping by use case is also helpful. I don't need a "best overall" winner. I need to know which tool fits which problem. That's much more useful.
Saving this one. When I eventually need a portfolio site for my writing, I'll come back to this list.
Thanks for the kind words, this is a good reminder that the "second version" is where these tools really get tested. I'd add one thing for anyone starting out: the first draft is the easy part, and learning to make a small change without breaking what's already there is a skill in itself.
Since you're writing tutorials, B12 might be worth a look for a portfolio site. It's built for quick first drafts, and the AI chat plus visual editor makes it easy to adjust things without writing code. Your prompt will matter a lot, so spend a few minutes describing your audience and what you want visitors to do, like read tutorials or contact you.
I'd also suggest testing the "boring edit" from the post early: ask for a change like moving a section or adjusting a heading, then ask for another change on top of it. That will show you quickly whether a tool is worth committing to.
Grouping the tools by what they're actually strong at, rather than ranking them 1 to 8, is the right call. "AI website builder" covers two different products: template-and-copy generators (B12, Wix, Squarespace) and code-generation environments (Replit, v0, Lovable, Base44). Comparing them head to head isn't very useful.
A few things I'd add for anyone choosing:
Your last point is the key one: AI shortens the time to a first version, but it doesn't remove review, testing, and maintenance.
Curious whether you hit any limits with custom domains or form handling on the free tiers?
The point about two different product categories is exactly why I grouped the tools instead of ranking them.
You're right on all four points. Lock-in and credit-based pricing are the two that surprise people most. The sticker price rarely matches what a project costs once you're iterating heavily, so I'd always budget extra for the first week. And I agree that generated auth, validation, and permissions need a manual review before launch, especially on the tools that build the backend for you.
On your question, custom domains were the most consistent limit. On most of these, the free tier gets you a subdomain and a preview, and the domain is where the paid plan starts (Base44, Wix, and Squarespace all work this way). Forms were more mixed, and they were the thing I'd test first on any free plan before committing.
Also, thanks for catching the table order. It doesn't match the numbered list, so I'll fix that.
I like how you tested every builder against the same five questions before ranking anything. Asking what happens after the first draft is the right test, since a site that looks good on day one can still hit a wall once you need to edit it, add forms, or move it to your own domain.
The split between fast business-site tools and app-oriented builders is the part I found most useful. It makes the list easier to act on, because readers can match the category to their project instead of comparing eight tools on the same terms.
Have you run into a builder that handled the custom domain step cleanly? In my experience that's where many of the easy demos start to fall apart, so your take on it would be interesting.
Your framing of AI builders as either quick business-site tools or app-oriented platforms helped me sort the list in my head. I especially liked that you judged each one on what happens after the first draft, not just how the initial page looks.
Which of the eight would you pick first for a small service business that only needs a contact form and a booking page?