CATCHES Launches RealFit AI Sizing Technology for Personalized Fashion Try-Ons

CATCHES, the London/New York physics-backed AI company, debuted its RealFit technology at NVIDIA GTC 2026 in San Jose, with the first public deployment going live on the website of California luxury label AMIRI, founded by Mike Amiri in 2014 and long associated with Americana aesthetics and distressed denim.

RealFit works by asking shoppers to enter their measurements and submit a photograph, which the system uses to generate a digital twin. From there, they can preview any garment in a brand's collection on their own body, toggling between sizes to compare how each piece fits, moves, and drapes before placing an order. The Bangkok Post described the resulting twin as "millimetre-accurate," though that specific claim has not been independently verified beyond the brand's own characterization.

The technology is built on NVIDIA's CUDA platform, with an OpenUSD foundation that enables integration with tools including Blender and Unreal Engine. RTX ray tracing, DLSS, and DLAA contribute the rendering quality that CATCHES describes as "mirror-like realism."

The business case Ed Voyce, Founder and CEO of CATCHES, is making is blunt: "Fit uncertainty lowers online conversions and causes over 50% return rates in some categories. CATCHES RealFit technology fixes this by providing customers a personalized photorealistic experience that understands real materials, cuts, and true sizing and fit." A Businesswire release attributed to the AMIRI brand noted that "CATCHES RealFit embodies" the label's values of community, craft, and innovation, though the specific speaker was not identified in available materials.

CATCHES described itself as currently working with additional brands beyond AMIRI, with further launches scheduled across the coming months, and additional luxury partners expected to come online through the remainder of 2026. The company positions itself as building both the application and infrastructure layers for what it calls "the next generation of digital fashion experiences," and is classified as an independent software vendor for the fashion industry in its NVIDIA partnership context.

Whether RealFit delivers on the accuracy it claims at scale will be the real measure of its ambition. The return rate problem it targets is one the entire fashion e-commerce industry has tried and failed to solve for a decade; a physics simulation that actually accounts for how fabric behaves on a specific body, rather than simply mapping a texture onto an avatar, would represent a meaningful shift in how personalization functions at the point of purchase.