The gallery is big and the storefronts look finished. What none of the ten ecommerce ones describe is where an order goes once the card clears. Here is the checklist to run a starter against before you give it a weekend.
A starter is a head start or it is nothing. The only reason to open a gallery instead of a blank prompt is that someone has already made the calls you would otherwise spend the first day making. Which of those calls it made for you is the thing you are shopping for, and a screenshot answers almost none of it.
Lovable's gallery listed 192 projects on 5 August 2026, ten of them filed under ecommerce, and they are genuinely nice work. A fashion storefront the blurb describes as a working cart and stockists directory, a coffee order page listed as cart checkout with menu filtering, an audio product page built around a variant grid. If what you need is a shop that looks finished by Friday, remixing one of those beats starting from nothing, and it is not close.
Read the ten descriptions next to each other, though, and they are all describing the same half of the shop. Wishlists. Filter panels. Image galleries, collection grids, email capture. Every one of those is something a customer looks at. Not one of the ten mentions where the order goes after the card clears, or who signs in on Monday to find out what leaves the building.
That is not a knock on the gallery. A template is a front end because a front end is the part that can be handed to you whole. The rest of a shop has to know your own stock and your own prices, so it tends to arrive as homework instead.
The general version of this, which tool quietly leaves which job on your desk, is worked through in a look at what AI tools handle and what nothing handles yet.
The checklist below is the one worth running before you commit.
Here is the trap, and it is a quiet one. The template hands you a finished looking front end on day one, so the project feels almost done. Everything still missing is invisible, because none of it has a screen yet.
Then the first real order arrives and the missing half turns up all at once. A number has to come down somewhere. Someone has to be able to look up what a customer paid, and change their mind about it two days later without touching the database by hand. Refund it and every one of those has to move backwards.
You can prompt your way there. People do, and it works. It just turns the afternoon the template saved into a month of evenings, and this is the half where a mistake costs you a customer rather than a screenshot.
Connect GitHub, Netlify, and Supabase in one click each, and the whole shop runs on your own infrastructure.
Your shop is built on standard Next.js and your data lives in your own Supabase project.
Describe what you sell and who you sell it to.
Get it built by AI from front to back.
Your shop goes live ready to take its first order today.
If nothing on the site ever takes money, go and remix one. A lookbook, or a page that collects an email before a launch: two of the ten are already close to that, and MaShop would hand you a commerce model you have no use for.
The design gap deserves saying plainly too. A gallery that size is a real asset, the defaults look finished, and a template someone drew by hand will out-photograph a generated shop on day one. MaShop puts the working parts first, and on your first afternoon with it you can see that choice being made.
If you are weighing the two products rather than the two starting points, the longer version of that trade sits on replacing Lovable for a commerce project.
And if the shortlist still has Bolt on it, the head to head on Lovable and Bolt reads both vendors on their own docs and says which one suits which job.
If the templates were only your way into a bigger question about the product itself, the rest of it, what the payments feature is scoped to and where physical goods get routed, is read off the same documentation in our full review of Lovable.
For the plain version of the question, without a competitor in the frame at all, there is what an ecommerce website builder has to generate.
Built around what AI costs not what your shop earns.
Try MaShop on a side project. No card required.
Includes:
For people running a commerce app full-time.
Includes:
Built for agencies and high-volume teams.
Includes:
Ten were listed under the ecommerce category on 5 August 2026, out of 192 templates in the gallery overall. They include a bakery order page and a charity auction site, so the category is wider than a product grid and a cart.
You can get the storefront from one, and further than people assume: three of the ten describe a cart or a checkout flow already. What the blurbs do not describe is the half behind it, the stock count, the order record and somewhere to manage both. That part you prompt in yourself, whichever tool you started from.
Whether it already contains the parts you cannot see. A product page and a cart are easy to verify because they are on screen. Stock that moves on a sale, an order record and a back office are the ones that decide whether the template saved you a week or a day, and none of them show up in a screenshot.
No, and that is the trade. There is nothing to browse: you describe the shop and the builder generates the storefront and the admin around what you sell. That skips the step where you find a template close to your business and then bend it, and it costs you the reassurance of seeing the finished thing first.
Their published plans start at $25 a month for the Pro tier with 100 credits, and $50 for Business, read on 5 August 2026. Remixing a template from the gallery is a normal build, so it draws on the same credit balance as anything else you make.
One prompt, and the storefront and the back office are generated together.