Korva is a scroll of the internet’s best ideas.
Weekly outfit plans
The plan comes back as pictures of the user wearing each outfit, top to bottom, with the shoes matched to the top for that day. Every week of outfits uses the clothes already in the closet.
Shopping from scans
Shopping runs on the same scans. The assistant searches the internet and returns thousands of outfits from many stores, shown on the user's own body. The user points to one and asks for it to be sent to their house, and the order goes through.
Some users ask for three new head-to-toe outfits every day. One click on a favorite buys it.
How can an app show clothes on your body?
Virtual try-on tools build a model of a person's body from photos or scans, then render a garment so that it follows that shape. Modern versions use diffusion models, the same family behind many AI image generators, to fold and drape fabric realistically. Google introduced a diffusion-based virtual try-on feature for apparel shopping in 2023. The tools were trained on pairs of garment images and photos of people in many poses.
Why do clothing purchases come back so often?
Fit is among the most cited reasons shoppers send apparel back, because sizes differ between brands and a screen cannot show how fabric hangs. The National Retail Federation and Appriss Retail estimated that returns in the United States reached 743 billion dollars in 2023, about 14.5 percent of sales. Online purchases are returned at a higher rate than store purchases. Retailers have invested in size tools and virtual try-on to reduce that cost.
How can software tell what clothes you own?
Computer vision models can label images by type, color, pattern, and style. A model called CLIP, released by OpenAI in 2021, learned to match images with text descriptions, which lets software sort photos by phrases such as striped cotton shirt. Once each item has labels, outfit planning becomes a search for combinations that match.
More from Invest Like the Best with Patrick O'Shaughnessy
