Kasane

• 

  One new item

• 

  Four different looks from one new item

• 

  creator settings

• 

• 

  gap hero

• 

  my wardrobe home

• 

  Open the wardrobe

• 

  wardrobe scan

• 

  video for reel/insta story gen

• 

  retailer handoff

Inspiration

Most people do not need more clothes—they need to understand what belongs in the wardrobe they already own.

We were inspired by a familiar frustration: finding something beautiful online but not knowing whether it will work with anything at home. This problem is even sharper for Instagram creators, who need outfits that can move across different settings and content formats without constantly buying more.

We wanted to answer one focused question:

What is the single missing piece that would make the rest of your wardrobe more useful?

What it does

Kasane studies the colours and garment categories already inside a person’s wardrobe. It compares them against Sanzo Wada’s historic colour combinations and identifies the highest-value wardrobe gap: the one new item that unlocks the greatest number of useful outfits.

In our demonstration, Kasane discovers an oxblood wide-leg trouser that works with nine pieces Jo already owns and creates four distinct looks.

YouCam then transforms that recommendation into visual proof:

YouCam Clothes V4 previews the garment on the shopper.

YouCam Background Change places each look into an appropriate setting: creative office, casual café, coastal vacation and gallery night.

YouCam Video V2 turns the four settings into five-second, creator-ready videos.

A verified retailer link provides a real path to purchase, with current price, size and stock confirmed at the retailer.

One wardrobe gap becomes four outfits, four settings and twenty seconds of content.

How we built it

Kasane is a React and TypeScript application supported by a dedicated colour-science and wardrobe-matching engine.

The engine:

Reads the colours and categories of owned garments.

Measures perceptual colour distance.

Tests those garments against a working collection of Wada combinations.

Uses distinct-garment matching so one item cannot incorrectly fill multiple outfit roles.

Ranks missing items by the number and usefulness of the outfits they unlock.

We designed the product in two connected layers. The first is a digital wardrobe containing the user’s existing clothes. The second is the Kasane recommendation, which clearly shows how the proposed purchase was discovered from that wardrobe.

YouCam requests are handled server-side so credentials never reach the browser. The integration uploads approved images, creates asynchronous YouCam tasks, checks their status and preserves successful outputs before their temporary URLs expire.

Challenges we ran into

The hardest challenge was making the recommendation genuinely useful rather than merely attractive. Early versions could create impossible outfits, reuse one garment for multiple roles or overstate the strength of a colour match. We introduced distinct garment assignment, clearer scoring and explicit limitations.

Lower-body virtual try-on brought another constraint: a standalone trouser packshot was not sufficient as the reference. We needed an appropriate worn-garment reference while clearly describing the result as an appearance preview—not a guarantee of size or fit.

YouCam’s background and video APIs are asynchronous, billed operations with temporary result URLs. We created a controlled generation workflow, saved verified outputs and prevented the public demo from becoming an unlimited credit-consuming endpoint.

Finding a real shopping path was also difficult. Most product APIs require commercial or affiliate approval, while retailer pages can change price and availability. We therefore show a dated, closest live product match and send the shopper to the retailer to verify current details instead of pretending to maintain a fictional cart.

Accomplishments that we're proud of

We are proud that Kasane is more than a visual concept. It is a complete, working journey from owned wardrobe to purchase decision.

Our biggest accomplishments include:

Turning a varied digital wardrobe into a specific, explainable recommendation.

Combining historic Wada colour science with modern wardrobe matching.

Demonstrating three YouCam capabilities in one natural product flow.

Producing four verified outfit settings and four five-second creator videos.

Preserving the original wardrobe experience so the recommendation feels grounded in real possessions.

Connecting the recommendation to a real retailer instead of ending with a fake cart.

Building a mobile-first public demo and publishing the complete architecture and source code.

Most importantly, every generative step supports a real decision instead of existing as an isolated AI effect.

What we learned

We learned that generative AI becomes far more valuable when it follows a strong decision system.

Kasane decides what deserves to be bought. YouCam makes that decision visible, contextual and shareable.

We also learned that context is an important part of purchase confidence. Seeing an outfit on yourself is useful; seeing it work at the office, on vacation and at an evening event is considerably more persuasive.

Finally, we learned that meaningful personalization does not need to begin with another shopping feed. It can begin with the clothes a person already owns.

What's next for Kasane

Next, we want to make wardrobe onboarding nearly effortless by allowing people to upload closet photographs, selfies or an existing wardrobe grid.

We also plan to:

Expand and audit the complete Wada colour dataset.

Add personal style, budget, sizing and fabric preferences.

Connect to approved retailer and affiliate product feeds for live discovery.

Introduce more advanced fit and measurement guidance alongside virtual try-on.

Generate complete creator kits with Reel covers, captions, hooks and shoppable links.

Learn from saved, rejected and purchased recommendations.

Create creator storefronts where every promoted product can be explained through the outfits it unlocks.

Our long-term vision is a shopping system that rewards usefulness rather than volume.

Kasane: The science of fashion, starting with your own wardrobe—buy less, unlock more.

Built With

• openai
• react
• restapi
• typescript
• youcam