AI-powered aggregator of clothing for Muslim women

If you’re a Muslim woman, your clothing has to meet a number of requirements: it shouldn’t be transparent or tight, it should look modest, and it should cover everything except the face and hands.

Modern fashion marketplaces contain thousands or even millions of items. For Muslim women, choosing suitable clothing is more challenging: they have to carefully review each product page to determine whether it meets these requirements.

Our client set out to change this.

A woman visits Scanjab and opens the catalog. There she finds items from ASOS and Farfetch — each product listed has already been checked for compliance with Muslim modesty criteria.

To find what she needs, she can use the search bar or choose a category such as dresses, shoes, headwear, etc. She can also adjust filters by color, brand, and price. Items can be sorted by novelty, cost, or popularity.

When she finds the right item, she clicks “Buy” and is redirected to the product page on the marketplace.

Once a day, our script searches for new items on ASOS and Farfetch through their APIs. After that, an LLM selects the products that are suitable for Muslim women.

We trained the model to identify appropriate items among new arrivals. It filters results using dozens of keywords — for example: modest, long sleeve, loose fit.

Additionally, the AI determines the color of the item based on its photo — this is necessary for the color filter to work correctly.

Each item on Scanjab includes a referral link leading to its page on the marketplace. Users follow the link and purchase the item; through the affiliate program, the client receives a percentage of each sale.

A flexible admin panel allows changing the criteria for selecting items at any time. For example, if a new tag is added and some items no longer meet the requirements, they are automatically removed from Scanjab.

The administrator can also edit product categories and modify the filtering system.