Okto AIOkto · wardrobe app

AI that knowsyour wardrobe

Photograph a piece — a model reads its color, fabric, cut and brand, lifts it off the background and files it away. Another one dresses you from what you already own, and answers in the chat with clothes instead of advice.

Download for

iOS 18+ · open this page on the device

The wardrobe screen in Okto
01

Context

Problem

The clothes exist, the data does not

People forget what is already in the wardrobe and buy the same thing twice. Nobody keeps a spreadsheet of their clothes — typing it all in never pays off.

Approach
Okto AI

The model does the typing

The only thing a person does is take the picture. A vision model fills in the rest — category, color, fabric, pattern, brand, styles, occasions — and a second one redraws the shot as a clean catalog image.

Status

Built and running

The app works end to end today: batch capture, a filtered wardrobe, an outfit editor, a chat assistant, and two languages with a fully mirrored interface.

02

Inside the app

The add menu: camera, photo library, new outfit
Adding

Shoot it, or take it from the camera roll

A piece gets in either way: straight from the camera, or picked from photos you already have — several at once. Both go through the same recognition, so the wardrobe stays uniform.

The same menu starts a new outfit, so building one never means hunting for a button on another screen.

Wardrobe

Everything you own, in one grid

Pieces sit cut out on a clean background, so they read the same way whatever the room they were shot in looked like.

Scroll through everything at once, or narrow it down to a category. What used to live scattered across a camera roll finally sits in one place.

Filters

Find it the way you think about it

Filter by color, brand or style — or by category from the row on top. The values come from what is actually in the wardrobe, so nothing in the list leads to an empty screen.

An item card with recognized attributes
Item card Okto AI

Everything the app read, in plain sight

Category, length, sleeve, material, color, styles and occasions — all filled in by recognition, all editable. The badge marks a shot the app cleaned up; the original is kept underneath.

Canvas

Laid out by hand

Add pieces from the shelf below, drag them, turn and resize them with your fingers; the selected one carries its actions right above it.

Or Arrange sorts the canvas out on its own — top above bottom, shoes at the floor, bag to the side, the way clothes are actually worn.

Assistant Okto AI

It answers with clothes, not paragraphs

It sees the whole wardrobe, the outfits already saved, and what you told it about yourself — so it never asks twice about the same thing.

A proposed outfit arrives as cards inside the chat: the person stays in the conversation instead of being sent off to browse a catalog.

The wardrobe in Arabic with a mirrored layout
Languages

Arabic is not a translation layer

Switching the language mirrors the whole interface: the profile moves to the right, categories run right to left, the tab bar flips. Colors, styles and occasions are translated too, so nothing stays half English.

The language is a setting inside the app, not the phone's — and it applies instantly, without a restart.

03

How it works

Step 1

Shoot a batch

The camera stays open after each shot, so an entire shelf goes in one go. You keep shooting while everything already taken is being sorted out.

Step 2

It reads the piece

Color, fabric, cut, brand — everything comes back already filled in, and everything stays yours to correct. Nothing is decided behind your back.

Step 3

It looks like a catalog

The piece ends up on a clean background, without a body and without the shadows of your room — the way it would look in a shop. Your original photo is never lost.

Okto AI

AI turns a camera roll into a catalog

  • A structured catalog of clothes instead of a camera roll
  • Advice drawn from what is already owned, not from a shop window
  • Wear history: what sits untouched, what goes on every week
  • A ready basis for personalization: styles, occasions, preferences
04

Where the intelligence sits

Native

A native iOS app that feels like one: instant, gesture-driven, with the wardrobe stored on the device and available with or without a connection.

On-device

Cutting a piece out of its background happens on the phone itself — nothing is uploaded to see the shape of a garment, and it works the moment the shutter closes.

Recognition

What the clothes are, and how they should look on a clean background, comes from generative models behind our own service — replaceable as the field moves, without touching the app.

Conversation

The assistant does not just talk: it shows pieces, assembles outfits, saves them and opens what it mentions — and reaches for live facts like weather when the answer depends on them.