Perfect Corp’s AI clothes try-on is real, it works roughly the way the coverage says, and you cannot find out what it costs. The product page for its AI Fashion Try-On offers a web demo, a “Contact Sales” button and a demo request form, with no published price and no self-serve tier (perfectcorp.com). That is not a criticism of the company. It is the single most useful fact for a small clothing store, because it tells you which market you are actually shopping in.
Who the headline products are built for
The generative try-on tools that get written up, Perfect Corp’s included, are sold the way enterprise software is sold: a demo, a scoping call, a negotiated contract, an integration project. If you run a shop with a few hundred SKUs and no developer, that process is not a pricing problem, it is a fit problem. You would be buying a platform capability and paying for an implementation you do not have the team to run.
The same underlying capability, though, is now sold a second way, through app marketplaces, at prices published on the page and installable without a contract. That is the version worth your attention, and it is what the rest of this covers.
Start with returns, not conversions
Try-on is almost always pitched on conversion lift. For a small apparel store the more honest place to look is returns, because that is where the money leaks quietly and where fit uncertainty is the actual cause.
The National Retail Federation, with Happy Returns, put the 2025 figures at 15.8 percent of annual retail sales returned, totaling 849.9 billion dollars, and an estimated 19.3 percent of online sales returned (nrf.com). Roughly one online order in five comes back across retail as a whole, and apparel sits at the harder end of that because sizing is the thing people cannot check.
Here is the discipline that matters more than any vendor’s claim: nobody can tell you what try-on will do to your return rate. Vendors publish improvement figures from their own customers. Treat those as marketing until you have your own number. What you can do before spending anything is establish the baseline, and that takes an afternoon: pull your last 100 online orders, count how many came back, and work out what each return actually costs you once return shipping, the payment processing you do not get back, and the time someone spends restocking are counted. That figure is what any subscription has to beat.
What it costs when you buy it the way a small store can
Two Shopify apps show the real shape of this market. Vue’s Virtual Try On runs 20 dollars a month on Starter, 35 on Professional and 75 on Business, each tier including a set number of try-ons and charging between 0.12 and 0.18 dollars for each one beyond it, with a 3-day free trial (apps.shopify.com). Antla’s AI Virtual Try On starts at 19.99 dollars a month on its Trend plan with 100 try-ons included and 0.16 dollars for each additional one, then 49.99 dollars for Runway with 500 included at 0.12 dollars over, and 199.99 dollars for Unlimited Fashion with 2,000 included at 0.09 dollars over, on a 7-day free trial (apps.shopify.com).
Note what is being metered. You are not paying per sale, you are paying per try-on session, and a try-on is something browsers do far more often than buyers do.
The tier arithmetic that catches people out
This is the part worth doing on paper before you install anything, and it is where the obvious choice is usually wrong.
Take Antla’s published numbers and run one volume through all three tiers. At 2,000 try-ons in a month, the cheapest-looking plan, Trend at 19.99 dollars, includes 100 and bills the other 1,900 at 0.16 dollars, which is 304 dollars of overage, for 323.99 dollars total. Runway at 49.99 dollars includes 500 and bills 1,500 at 0.12 dollars, which is 180 dollars, for 229.99 dollars total. Unlimited Fashion, the 199.99 dollar plan that looks like the expensive one, includes all 2,000 and costs 199.99 dollars.
The most expensive sticker price is the cheapest bill. That inversion is normal in metered pricing and it is why the entry tier is the one that surprises people: it is priced to be attractive at volumes you exceed in your first busy week. Before choosing, estimate try-ons rather than orders, and estimate them high.
The same reasoning applies to whether the subscription earns its place at all, which is the ordinary question of whether a channel pays for itself or just feels productive, and it depends on knowing your real per-order cost, which is the same baseline discipline as documenting what a unit actually costs you.
What try-on does not fix
It does not fix a size chart that is wrong, and it will not stop returns caused by a garment that genuinely runs small. A visualization of a product is only as honest as the product data behind it, and a rendering that flatters what arrives in the box produces a return plus a disappointed customer, which is worse than the return alone.
It also does not remove the person who knows the stock. In a small shop, somebody can tell a customer that a particular brand runs narrow, or that the linen relaxes after a wash, and no rendering carries that. The useful version of this tool takes the repetitive “will this fit me” traffic off that person’s plate so they can spend their time on the questions that actually close a sale and on the customers who need real advice. A store that installs try-on and then decides it no longer needs the person who knows the inventory has traded its one genuine advantage over a marketplace listing for a monthly subscription.
What to do before you install anything
Get your baseline return rate and your real cost per return on paper first, because without them a trial tells you nothing except whether the widget loads. Then run one free trial, on your highest-return category rather than your bestseller, and watch try-on session counts as closely as sales, since that is the number your bill is made of.
Decide the volume tier after the trial, not during signup. It is the one decision here that is easy to get wrong and dull to unwind.
Frequently Asked Questions
Can a small store actually buy Perfect Corp’s try-on?
Not off the shelf. Its AI Fashion Try-On page offers a web demo and a “Contact Sales” or demo request path, with no published pricing and no self-serve plan (perfectcorp.com). Small retailers generally reach the same capability through app-marketplace products that publish their prices.
How much will try-on reduce my returns?
Nobody can answer that for your store, and treat any vendor figure as marketing until you have measured your own. Establish your current return rate and cost per return before a trial starts so the comparison afterward means something.
Why did my bill come in higher than the plan price?
Because these apps meter try-on sessions, not sales, and charge for sessions past the included allowance. Vue lists overage between 0.12 and 0.18 dollars per try-on depending on tier (apps.shopify.com) and Antla lists 0.16, 0.12 and 0.09 dollars across its three plans (apps.shopify.com). Browsers generate try-ons whether or not they buy.
Which product category should I test it on first?
The one with your highest return rate, not your bestseller. Try-on is a fit-uncertainty tool, so testing it where fit uncertainty is worst gives you the clearest read, and a bestseller with low returns has little room to improve.
