AI try-on On-model photoshoots 360° video Live makeup AR

One product photo.
Four ways to sell it.

AI try-on on your product page, AI on-model photoshoots, 360° video and live makeup AR — all from the product photos you already have. One click on Shopify, one snippet anywhere else.

Ten images free, no card. A generation that fails never costs a credit.
Installs on Shopify in one click — no code edits, no reshoot.

A flat-lay photograph of an emerald satin dress with a black lace bodice, a gold belt, a clutch and heels.
Real The same emerald dress rendered on a model.
On-model
Real A shopper's own photograph with a lace gown rendered onto her.
On a shopper
Real
360° video
Interface A close portrait showing a lip colour and blush.
Live AR
How it works

Four things, one wallet

Every one of them runs off the product photos you already have, and draws from the same balance. Here is exactly what it looks like.

  1. A “Try it on” button appears on your product pageSelvette adds the button. The rest of the page stays your theme.
  2. They upload a photo, or take one there and thenThe camera opens in the browser, with a self-timer so they can step back for a full-length shot.
  3. They see the exact variant on themselves, then add it to the cartThe size and colour selected on the page is the one that gets tried on, and the one that goes in the bag.
Apparel · bridal · swim · modest wear See use cases →
your-store.myshopify.com/products/emerald-lace-dress
Maison Rue Emerald lace dress $189.00
SML
Add to cart
Try it on

Try it on

See how it looks on you
Take a photo
Upload a photo

Your photo is only used for this try-on and auto-deletes in 24h. We never train on it.

Try it on

See how it looks on you
Generating your try-on…

Wearing it

Add to cart · M
Fidelity

The hard ones

Anything looks fine on a plain tee. The test is a garment with structure: black lace over emerald satin, a gold buckle, a high-low hem. Drag to see what went in and what came back.

Flat-lay of an emerald satin dress with a black lace bodice, gold belt, clutch and heels. The same dress on a model, with the lace pattern, buckle and high-low hem preserved. Flat-lay in On-model out
Flat-lay of a navy swimsuit photographed on a plain ground. The same navy swimsuit rendered on a model. Flat-lay in On-model out
  • Buttons stay buttons.Designed to preserve collars, hems, sleeve length and hardware. A garment that comes back subtly redesigned is a recurring complaint across this category.
  • The person stays the person.Face, hair, skin tone, apparent age and any head covering are preserved, and modest coverage is respected rather than quietly removed.
  • Nothing here is retouched.Both frames came out of the pipeline this product runs on.
The fine print, up front

Built for the parts AI usually gets wrong

Three things merchants complain about across this whole category. We designed around them rather than hoping you would not notice.

A failed generation costs nothing

Credit is reserved when a job starts and released if it fails, errors or is rejected. Nothing is deducted for work we did not deliver.

No bring-your-own-key

Model improvements land on the plan you already pay for. You will not be sent off to buy an API key to use something you already bought.

It never blocks a sale

If you run out of quota the try-on button is removed, not shown broken. Your shopper never sees an error, and never sees a billing state.

The shoot you didn't have to book

Most of what a product shoot costs is coordination — the studio, the model, the stylist, the day. A new colourway means doing all of it again. Here it is a regenerate.

It works off Shopify too

One click on Shopify. On WooCommerce, a headless build or your own site, it is a single script tag and a product id — the same widget, the same wallet.

It stays light

The bundle loads lazily, the AR engine only downloads when someone taps it, and nothing is injected into your checkout.

Insights

You can see exactly what it did

Most apps in this category report that people clicked. This one follows the shopper through to the order, and hands you the emails it collected on the way.

Insights · last 30 days Sample data
Saw the try-on button
4,812
Opened it
1,974
Generated a try-on
1,268
Added to cart
529
Ordered
287
LeadProductConsent
r••••@gmail.comEmerald lace dressYes
a••••@outlook.comSatin lip colourYes
m••••@icloud.comNavy swimsuitYes

Revenue, not sessions

Try-on to add-to-cart to order, with orders and revenue attributed back to the shoppers who tried something on — shown against your store's unattributed orders so the number means something.

Every try-on can be an email

Optional capture, before or after the result, consent recorded per lead, CSV export. Off by default — you decide whether it appears at all.

Per product, not just per store

Which products get tried on, which convert, and which are burning generations without selling anything.

Pricing

One wallet, four capabilities

Images and video draw from one balance, shared across every store you connect. Each store carries its own plan, and every plan includes a monthly allowance of live AR views.

Free

$0
  • 10 credits, one time
  • 10 live AR views
  • No 360° video
  • = 10 AI images
Start free

Grow

$19/mo
  • 150 credits / mo
  • 1,000 AR views / mo
  • No 360° video
  • Billed per store
Choose Grow

Starter

$59/mo
  • 500 credits / mo
  • 5,000 AR views / mo
  • No 360° video
  • Billed per store
Choose Starter

Scale

$249/mo
  • 2,600 credits / mo
  • Unlimited AR views
  • 360° video (10 credits)
  • Billed per store
Choose Scale

Out of credits mid-month? Optional extra usage keeps generating — 13¢–9¢ per image by plan (360° video $1.00–$0.90, Pro and above) — off by default, with a monthly cap you set.

Running Selvette on your own website, or managing client sites as an agency? Talk to us for a plan that fits.

Questions

The things people get burned on

Every question below is one that merchants have publicly complained about somewhere else in this category.

What happens when a generation fails?

You are not charged. The credit is reserved when the job starts and released if it fails, errors, or is rejected before delivery.

How do the credits work?

One AI image is one credit; a 360° video is ten. A monthly plan's allowance resets on your billing anniversary — the day you subscribed, not the first of the month — and is shared across every store you connect. The free plan's credits are one-time and never expire. Live AR doesn't spend credits at all: each plan includes its own monthly allowance of AR views. And if you run dry, optional extra usage keeps generations flowing at a per-image price with a monthly cap you control.

What is the difference between AI try-on and live AR?

AI try-on is for garments: the shopper gives a photo and the piece is generated onto them. Live AR is for makeup: lips, blush and eyeshadow rendered on their camera feed, on their own device, with nothing uploaded and nothing generated.

What happens to a shopper's photo?

It is used to produce that shopper's try-on and nothing else. Results expire on a retention window you set per store.

Will it work on my theme?

It installs as a theme app embed with no code edits. The button renders inline under your buy buttons where the theme allows it, and falls back to a floating trigger where it does not. It never blocks Add to cart.

Do I need to reshoot my catalogue?

No. It works from the product photos already in your admin.

What happens when a better model ships?

Improvements land on the plan you already pay for. If a future model ever costs materially more to run, it will be priced openly and announced before it ships, not discovered afterwards. You will not be asked to bring your own API key.

Does the shopper need an account?

No. They upload, generate and add to cart. Email capture is optional, and you decide whether it appears before or after the result.

Put your catalogue on a real person.

Ten images free. No card, no call, no onboarding project.