Open‑Source Wardrobe Skill: Let Codex Organize Your Girlfriend’s Outfits

Convert Photo Album to a Digital Wardrobe

Wardrobe is an open‑source project that uses AI to extract individual clothing items from a regular outfit photograph (top, bottom, shoes together are acceptable). The workflow is:

Detect all garments in the image.

Use the Images API to generate clean, transparent item pictures.

Store the original photos, generated images, and a local wardrobe database.

Classify items into tops, outerwear, bottoms, accessories, and shoes.

Display a tidy gallery of single items via a web UI, eliminating the need to browse thousands of phone photos.

Each garment receives a data card where users can edit the name, category, primary and secondary colors, and add descriptive tags (e.g., "casual", "vintage"). The interface extracts a palette of candidate colors from the item image, allowing manual selection or re‑sampling directly from the picture.

Codex Skills

import‑clothes – Point the skill to a folder of photos. It scans all images, lists detected garments, removes duplicates, crops reference regions, generates transparent item images, and saves only garments that pass a verification step.

generate‑outfits – Specify the desired number of outfits and optionally season, occasion, and style. Each generated outfit includes at least a top and a bottom, with optional outerwear, shoes, or accessories. The skill renders a Lookbook image by compositing the complete outfit onto a personal reference photo.

Installation and Usage

Requirements: Node.js 22 or higher.

git clone https://github.com/tandpfun/wardrobe.git
cd wardrobe
npm install
cp .env.example .env
npm run dev

Add your OpenAI API key to .env: OPENAI_API_KEY=YOUR_API_KEY Place a clear personal PNG reference image at data/model-reference.png. Open http://localhost:5173 in a browser to access the Wardrobe UI. Without an API key the AI‑driven import functions remain disabled, but the UI still loads.

Metrics

The repository gained over 1,500 stars on GitHub within two weeks of release.

Repository: https://github.com/tandpfun/wardrobe