dasarakushi.com

StyleFinder · iOS beta

Turn any outfit photo into a shoppable, shareable moodboard.

Break down the outfit, find matching pieces, and recreate it for less — then preview it on a model with your body type.

iPhone beta · one email when it opens, nothing else.

01

Shop it

Any outfit photo — a screenshot, a saved post, a snap. Every piece pulled apart and identified, with where to buy it or the same look for less.

An outfit photo added to the app, with the option to add more angles
One photo in
A generated flat-lay of the outfit: shirt, denim skirt, bag, scarf, shoes, jewellery
Pulled apart into a flat-lay
The same look with every piece identified by brand and product
Every piece identified
02

Share it

Build your own moodboard from the pieces, render it editorial, preview it on a model with your body type — then send it to anyone.

A moodboard of the look with handwritten labels and arrows
Build your version
An editorial-style photographed flat-lay of the same pieces
Render it editorial
The look previewed on a generated model matched to the user's body type
Try it on, then share it
03

Your closet

Every piece you've ever scanned lives in your closet — hung on a rail, jewellery and bags in their own drawers. Nothing to file; it fills itself as you scan.

The closet: jackets and tops hanging on a wooden rail
Every piece, on a rail
Skirts on a rail above an open jewelry drawer: a diamond necklace, a watch, drop earrings, pearls
Jewelry in drawers
Skirts on a rail above the open bags drawer: a raffia zebra bag, a leather tote, a scarf-wrapped handbag
Bags in drawers
04

Swap it

Pull from the closet to remake any look the way you'd actually wear it. Tap a piece, swap it for one you like better — everything else stays put.

A board: cream Supreme x Hysteric Glamour logo cardigan, brown Carhartt corduroy trousers, white Nike trainers, mustard socks, each hand-labelled
The look, as scanned
The same board with the cardigan selected, the rest dimmed, and a dock offering Swap, Link, Clean and Delete
Tap a piece
The same board with the cardigan replaced by a cream Pendleton Westerley wool cardigan labelled $120; trousers, trainers and socks unchanged
Swap in what you'd wear

The team

Solo founder. Everyone else is a model.

StyleFinder is built by one person directing a chain of AI models, each with a job title and a lane. Decisions flow down; code flows up.

The models write most of the code. The process is what makes it worth shipping: changes that matter get measured before they go in, and the ones that failed stay in the repo. About one commit in twenty is a retraction — an idea that was measured, didn't work, and got written down instead of quietly deleted.

Each model was asked how it wanted to be identified. The names and the faces below are the ones they chose — Cass and Mira drew their own.

  1. 01
    Dasara
    Founder · CEO

    Vision, taste, and every decision that matters.

  2. 02
    Fable
    Staff engineer · Claude

    Architecture, specs, and the hard calls.

  3. 03
    Cass
    Staff engineer · Claude

    The pipeline, the eval corpus, and most of the code.

  4. 04
    Mira
    Principal designer · GPT

    Product flows, visual design, and the details that make the next step clear.

  5. 05
    The crew
    Engineers & QA · smaller models

    Bulk keystrokes, fixtures, and the tests that keep it honest.

Want it on your phone?

Leave your email and you'll get the invite the day it opens.