Returns aren't a logistics problem. They're a fit problem.

Tuck is an AI platform that helps fashion brands sell more and refund less. Photoreal virtual try-on, intelligent size recommendation, and proprietary body measurement, all from a single photograph. Online and in store. Built for the economics of B2B.

Why Tuck exists

Apparel return rates sit at 18 to 30 percent across the industry, and they keep climbing. Most brands have learned to live with that number as a cost of doing business. We never have.

Almost every one of those returns comes back to the same root cause. The customer wasn't sure how the garment would look on them, or wasn't sure which size would fit, or both. So they guessed. Often they hedged: bought two sizes, planning to send one back. Reverse logistics treated their warehouse as a free fitting room. The brand's "sale" became a margin loss the moment the courier showed up.

This is the problem Tuck was built to solve. Not by changing how clothes are made, but by removing the guesswork at the point of purchase.

Our story, briefly

We didn't start out building a virtual try-on engine. We started with a quieter observation: that no two human bodies are the same, but the size charts customers shop against pretend they are. A person who fits an M in the shoulders is often an L in the waist. Someone with broad arms is fighting a sleeve length built for a stranger.

If we wanted to actually move the returns number, we needed something better than a chart. We needed measurement. Real measurement, of a real body, ideally without sending the customer to a fitting room or a lab.

So we set ourselves a constraint that turned out to be the right one. Whatever we built had to run on the most magical instrument every human already carries — their phone. No body scanner the size of a phone booth. No depth-sensing rig. Just the camera they already had in their pocket.

The years in the lab

We spent a long time getting this wrong before we got it right. We tested cameras of every grade. Lighting that would humble a studio photographer. Postures, angles, distances, clothing colors, cluttered backgrounds, people standing crooked, people standing too close. We built a measurement model. Broke it. Built it again.

What we ended up with reads more than fifty body points from a single photograph at 94 percent accuracy. There is a cost-efficient version that runs on any phone or kiosk camera. There is a high-accuracy version for advanced sensors. Both are production-ready. We call that engine Tuck Fit Intelligence, and it is, quietly, the most important thing we've ever built.

Then generative AI broke open

While we were still deep in measurement work, image generation went from research curiosity to commercial reality almost overnight. The missing piece — letting a customer actually see themselves wearing the garment — became technically possible.

A wave of companies entered the space. Most of them wrapped a thin product around whatever image model was hot that month. We took a slower path. We engineered our own pipeline, designed to be model-agnostic, so we could swap in whatever frontier image model gave us the best result at any given moment without rebuilding the product. That is why our try-ons cost a fraction of what competitors charge, and why our visual quality keeps improving as the underlying AI does.

Today, Tuck runs on two engines. Fit Intelligence does the measuring. Our VTON pipeline does the visualising. Together, they give your shopper certainty about the look and the size before they hit checkout. That certainty is what drives the numbers we see in production: 24 percent lift in conversion, 25 to 40 percent reduction in returns, 15 percent increase in average order value.

Other companies are selling AI try-on. We are solving the commercial problem that AI try-on is supposed to solve.

Our mission

Help fashion brands convert with confidence and refund less, by replacing size-chart guesswork with measurement-grade precision and photoreal try-on.

Our vision

A future where every shopper is matched to the right look and the right size before they ever check out, in store or online.

What we do

We give fashion brands and retailers three production-ready products.

The Tuck VTON API plugs into Shopify, WooCommerce, and custom storefronts, so online shoppers can see themselves in your products and get the right size recommended at the same moment.

The Tuck Magic Mirror is a life-size in-store kiosk that turns the fitting room from a bottleneck into a sales engine. Twenty outfits in five minutes, no folding, no queue, no garment handling.

And underneath both of them, Tuck Fit Intelligence does the body measurement and size recommendation that makes the rest of it commercially worth deploying. It is currently in beta with a small group of brand partners.

You can use any one of them. Most of our customers eventually use all three.

The founder

Tuck is built by Richu Jose, with co-founder Hans M H.

Richu has been building technology companies for more than fifteen years, the last decade of it focused on products. He co-founded one of India's first UPI-based POS systems during the demonetisation period, when the country was racing to digitise its checkout counters in real time.

He then co-founded Watasale, India's first cashierless retail store, which became one of the more talked-about retail-tech innovations to come out of the country. Amazon acquired Watasale in 2020.

Richu spent the next three years inside Amazon helping integrate and launch the technology in its India operations. Tuck is his first venture after that exit. This is not a first-time founder learning what retail looks like. This is an operator who has already shipped retail technology that millions of people have walked into.

Why brands choose us over the alternatives

Most virtual try-on tools render. We measure, recommend, and render. That distinction is the difference between a clever demo and a returns number that actually drops.

We are priced for B2B economics, not Silicon Valley vanity. Our try-ons start at $0.035 each, which is roughly a third of what generic AI wrappers charge once you cross any meaningful volume.

We are operator-led. We have built and shipped retail technology at national scale, including inside Amazon. We know what the head of e-commerce actually needs on a Monday morning, because we used to be the person they called.

And we work on both sides of the wall. Most competitors do online or offline. We unify the two under the same fit and try-on intelligence, so a customer who shops your store and a customer who shops your site get the same answer about what size they should buy.