WEARFITS virtual try-on: a six-month review of the Team’s work

Written by WEARFITS Team | Jul 15, 2026 10:14:06 AM

The past six months at WEARFITS have been a period of intensive technological development. Our focus has been on transforming Virtual Try-On (VTO) technology from an impressive novelty into stable, scalable infrastructure for modern commerce.

Our work has centered on three pillars: sizing accuracy, photorealistic visualization, and the automation of implementation processes.

Sizing, solved: addressing the footwear sizing challenge

June 2026 marked a major milestone in our history, as we officially launched the WEARFITS Virtual Try-On & Sizing module for Shoes.

For two years, we analyzed the needs of footwear brands, which repeatedly told us that visualization alone was only half the solution. The other half was answering the question: “Which size should I buy?”

Our solution combines augmented reality with precise fit analytics in a single browser-based experience.

We introduced a pressure heatmap that shows users, in real time, where a shoe feels tight and where it fits comfortably. A key UX improvement is the side-by-side size flick, which allows users to instantly compare the fit of two adjacent sizes without leaving the virtual try-on.

The entire experience is based on a 60-second foot scan performed with a smartphone. The scan is permanently saved within the WEARFITS ecosystem, eliminating the need to repeat the process for future purchases.

Gaussian Splatting: a new era of volumetric visualization

In July 2026, we became the first company on the market to introduce a virtual try-on solution for footwear and handbags based on Gaussian Splatting technology.

Traditional polygon meshes often struggle to accurately reproduce details such as the texture of silk, the sheen of leather, or material transparency.

Gaussian Splatting represents objects using millions of overlapping points of color and light, delivering:

Unprecedented realism: Accurate reproduction of fabric draping, light reflections on leather, and other details like beads.

Faster production: A reduction in the time required to create digital product twins from several days to just a few hours.

Accessibility: Models are rendered directly in standard web browsers without delays, which is critical for mobile conversion.

Gen AI for Apparel: precision over image generation

At the same time, we continued developing our Gen AI engine for apparel. In an industry where most tools rely on basic diffusion models, we focused on five technical areas designed to ensure credibility and consistency:

Preserving garment structure: Our model is trained to understand the construction of garments such as blazers and jackets, preventing the hallucination or distortion of details such as lapels and buttons.

Identity consistency: We ensure that after a garment is applied, the customer retains the same skin tone, body proportions, and distinctive personal features.

Fabric fidelity: The engine natively distinguishes between materials, allowing cotton to behave differently from satin.

API scalability: We built infrastructure capable of supporting thousands of SKUs while maintaining consistent quality and low latency.


2D-to-3D automation and e-commerce integration

We recognized that for the 42,353 footwear stores operating on Shopify, content creation is a major bottleneck.

To address this, we implemented 2D-to-3D automation that generates AR-ready models directly from standard product photographs. As a result, a brand can add 100 products to its virtual fitting room in just two hours.

Our Shopify and WooCommerce plugins have been optimized for performance to minimize user drop-off caused by long loading times.

We also introduced advanced Perfect Placement technology, which automatically handles product scaling and alignment in relation to the user’s body. Improved occlusion and masking models ensure that virtual items interact naturally with the user’s silhouette.


Accessories and technical details

We also continued developing our handbag experience by introducing dynamic positioning.

Users can now switch between viewing a handbag in their hand and wearing it on their shoulder with a single click. This functionality is supported by a physical movement simulation system called Realistic Motion Simulation.

The past six months at WEARFITS have marked a transition from simply “showing products” to enabling customers to truly “experience” them.

We have focused on stability and eliminating friction from the purchasing journey, based on the belief that trust is built through precision, not merely through impressive demonstrations.

Our goal remains to make online fashion shopping as confident, interactive, and engaging as shopping in a physical store.