How we compare AI website builders
We list 13 AI website builders available to Australian businesses, and a published rate card exists for 13 providers. Where a vendor quotes instead, we show a quote rather than an invented figure.
Two corrections are applied to every published price before it reaches the table. The first is the billing basis: a headline that only exists if you commit and pay up front is not a monthly price, so we record what the vendor is really charging month to month. The second is currency: where a builder bills in US dollars we convert rather than reprint, because a US dollar figure is not what leaves your account.
Ranking is arithmetic on those corrected rates at the number of websites and editors you enter. Commission never moves a builder up the table, we do not sell prominence, and the order is whatever the arithmetic produces.
The advertised price and the true monthly rate on an AI website builder
The most useful thing to understand about this category is that the number on a vendor's own pricing page is frequently not a monthly number at all. Prepay discounts here are unusually deep, and they are the reason a straight comparison of headline figures across this category tells you almost nothing.
Two separate mechanisms produce the gap. The first is a term commitment. Hostinger's contract row reads None on monthly, up to 48 months to get the advertised rate, and WordPress.com's reads None on monthly, up to 3 years for the advertised rate. Those are pricing terms rather than fine print: the advertised rate is the multi-year rate, the month to month rate is higher, and the renewal after the discounted term is higher again.
The second is a pricing page that simply opens on the annual view. Wix's billing basis is recorded as Annual only and its contract as Annual, paid in full up front, which is why its figure behaves differently from most of the table. Where a builder genuinely bills by the month with nothing to unwind, the row says so plainly: Webflow's basis is True monthly with a contract of None on monthly, and Squarespace's contract is None, monthly or annual.
Currency is the third correction, and for an Australian buyer it is a real one. Wix's local pricing row reads Yes, genuine AUD including GST, Squarespace's Yes, genuine AUD rate card and Hostinger's Yes, genuine AUD on the Australian site. Lovable's reads No, US dollars only, as do Bolt.new's and v0's. A builder that bills in US dollars will change what it costs you without ever changing its price, and most Australian cards add a foreign transaction fee on top.
The practical rule is short. Compare monthly against monthly, and if you intend to prepay, price the renewal rather than the first term.
Site or app: the two kinds of AI website builder in this table
This table looks like one market and is really two, and choosing from the wrong half is a more expensive mistake than choosing the wrong price.
One half generates a website: pages, a template, a content system, a contact form. The other half generates a working application, with a database, user accounts and payments behind it. The row that separates them is the backend. Durable's reads No app backend (site plus basic CRM and invoicing), Squarespace's No, CMS and member areas (no app backend), Framer's No, CMS collections (no app backend) and Hostinger's No app backend (front-end site). Against those, Lovable's reads Full-stack: Supabase (Postgres) + auth, Base44's Yes, full-stack (database + auth + payments), Bolt.new's Yes, full-stack (database + auth) and v0's Yes, full-stack via Vercel.
The templates row tells you the same thing from the other direction. Wix lists 900+ and Squarespace 150+, design-led, while Bolt.new lists None, output is generated and Lovable None, output is generated. A generator has no gallery because there is nothing to browse: the output is written for you each time, which is powerful when you want something specific and unhelpful when you want something proven.
If you are a designer or working to a brand, the design import row is where to look. Framer's reads Figma import, Webflow's Figma import and v0's Figma + image import.
Be honest about which you need. If what your business wants is a site that says who you are, what you do, where you are and how to contact you, the app generators are the wrong purchase and you will pay for capability you never touch. If what you want is a booking system, a member portal or something with a login, a template builder will fight you for months and you will end up rebuilding.
Per site, banded or unlimited: three cost shapes in AI website builders
Nearly all of the pricing confusion in this category comes down to one question the plan cards do not answer clearly: how many websites does this subscription cover?
Per website. The plan buys one site, and a second site means a second subscription at the same rate. Squarespace lists 1 per subscription, Webflow 1 per site plan, Wix 1 per plan, WordPress.com 1 per subscription and Durable 1 per subscription.
Banded by a site cap. The plan covers a set number of sites, and you move up a rung when you pass it. 10Web lists 1 / 2 / 4 by plan and Hostinger 3 (Premium), 50 (Business).
Unlimited and flat. The subscription covers as many as you build. Lovable lists Unlimited on paid plans, Base44 Unlimited apps on paid plans (5 on Free) and Bolt.new Unlimited projects.
Those three shapes are why this table reorders itself. A sole trader with one website and a marketing consultant running five sites for clients get genuinely different answers out of the same rate cards, and the gap between the answers is larger than the gap between any two entry prices.
Editors are the second axis, and it is capped harder here than in most software. Framer's pricing model is Flat plan price plus per-editor charge and its team row reads Owner free, then A$40 per editor, hard ceiling of 10. Wix's reads 2 (Light) to 100 (Business Elite), hard caps and Squarespace's 2 (Basic), unlimited from Core, while WordPress.com's reads Unlimited on every plan, including free. Enter your real site count and your real editor count into the calculator, because the entry price is the floor of a ladder rather than a price.
Why a free AI website builder plan usually fails on the domain, not the caps
Free plans in most software categories are throttled: fewer records, fewer seats, fewer sends. Here the free plan usually does the whole job and then stops at one specific thing, and it is worth naming it precisely.
The free tiers publish a real, live website. Wix's reads Yes, publishes live on a wixsite.com subdomain with Wix ads and WordPress.com's Yes, publishes live on a wordpress.com subdomain with ads. Framer's reads Yes, but no custom domain, Bolt.new's Yes, publishes a working app with Bolt branding, no custom domain and Webflow's Yes, Starter plan on a webflow.io subdomain, 2 pages, no custom domain.
The common thread is the custom domain. Wix gates it at Paid plans only (Light and above), Framer at Basic and above, Webflow at Basic plan and above and Base44 at Builder plan and above.
A website on a vendor subdomain, usually with the vendor's own advertising on it, is not a business website. You cannot put it on an invoice next to your ABN with a straight face, you cannot run a matching email address on it, and it is the one asset in this whole exercise that is genuinely yours: a domain moves between builders, nothing else does. So a free plan here is a way to see whether you like the tool, not a way to run a business, and you should assume you are paying from the day you go live.
Two smaller points sit under that. Some builders have no free plan at all: Hostinger's row reads No, Squarespace's No, 14-day trial only and 10Web's No, free trial only. And where a domain is thrown in, it is usually thrown in once: Hostinger's custom domain row reads All plans, free for the first year and WordPress.com's reads Paid plans, free for a year on annual and longer terms, so budget for the registration from the second year.
Metered AI credits, and when an AI website builder subscription is only a floor
On part of this table the plan price is the whole price. 10Web's pricing model reads Flat plan price, Durable's Flat plan price, Squarespace's Flat plan price and Wix's Flat plan price.
On the rest, the AI is a separate meter and the subscription is a floor. Lovable's pricing model reads Base plan plus AI credit consumption, Bolt.new's Base plan plus AI token credit consumption, Base44's Base plan plus AI credit consumption and v0's Per-user plus AI credit consumption. Building with these is iterative by nature: you describe, look, and describe again, and each pass spends the meter. The plan you need is decided by how much rebuilding you do, which is a number you cannot know before you start.
The allowances are published, and they are not shaped the same way. Bolt.new's reads 1M tokens/mo (Free), from 10M/mo (Pro), Base44's 25 message credits/mo (Free) to 1,200 (Elite) and Emergent's 10/mo (Free), 100 (Standard), 750 (Pro). Tokens, message credits and monthly credits are different units, so a credit count on one builder tells you nothing about a credit count on another.
Worth being precise about one thing: a credit grant does not necessarily rise with the plan price on the same vendor. Lovable's credit row reads 100/mo on Pro and on Business, so the step up between those plans buys something other than more AI. Read the credit row per plan rather than assuming it scales.
Some builders do not publish a number at all. Wix's row reads Not published as a numeric limit, Webflow's Not published as a number and WordPress.com's Not published as a number. That is not evasive on a builder where AI assists an editor rather than doing the building, but it does mean you cannot budget from it, so treat the AI there as a feature of the plan and not as a meter.
Finally, which model runs it. Base44's choose-your-AI row reads Yes, model selector (Builder+), v0's Yes, all models (Plus and above) and Lovable's Yes, Claude tiers (Sonnet/Opus), while Wix's reads No, fixed AI. On the models themselves, Bolt.new lists Anthropic Claude, 10Web lists GPT-4o, Gemini and Claude (multi-model) and Wix lists OpenAI (ChatGPT) + DALL-E. A fixed proprietary model is not a fault, but it does mean the quality of your output changes when the vendor changes its model, and you have no say in when that happens.
Renting or owning: what you can take with you from an AI website builder
Hosting is bundled everywhere in this category, so the ownership question is not about the server. It is about whether you can take the site somewhere else.
The row that answers it is code export. Wix's reads No, Squarespace's No, Durable's No and Framer's No, but supports custom code. On the other side, 10Web's reads Yes, it is WordPress, WordPress.com's Yes, it is WordPress, Lovable's Yes, GitHub sync, Base44's Yes, two-way GitHub sync and Webflow's Yes, code export.
Understand what each side actually buys you. Where there is no export, leaving means rebuilding: your content and your domain come with you, your site does not, and a redesign you did not plan is the real switching cost. Where there is export, you get a repository, and with it the job of hosting, patching and deploying it yourself. v0's hosting row reads Yes, via Vercel and Emergent's Yes, private hosting on paid plans, which is convenient while you stay and is the thing you inherit when you go.
Day to day, the difference you will feel sooner is support, and timezones matter from Australia. Wix lists 24/7 chat and callback, Hostinger 24/7 chat and 10Web 24/7 chat, while Bolt.new lists Community and email, Lovable Community and email and Framer Email and community. A community forum is a fine answer to a question you can wait on, and a poor one when your checkout is down on a Saturday.
The honest advice is to match portability to how much the site is worth to you. For a service business whose website is a brochure, a hosted builder you cannot export from is a reasonable trade for never thinking about hosting again. If the site is the business, if a developer will take it over later, or if what you are really building is an application, the export row deserves to outrank the price row.