Cross Cleanse

Skin Quiz
Personalized Routine
Shop
Homepage
Homepage

Problem Statement: Skincare is a $200B industry, yet most people are still guessing when it comes to their products. Many just buy products based on influencer recommendations or reviews, without understanding whether those products actually suit their skin type, address their specific concerns, or even work safely together. Dangerous ingredient combinations such as retinol + AHA/BHA are common knowledge among dermatologists but invisible to everyday consumers, resulting in wasted money, damaged skin barriers, and routines that don't work.

Solution Overview: Cross Cleanse is a web application that has users go through a focused 6-step skin quiz, then uses a content-based recommendation engine to generate a fully sequenced AM and PM skincare routine matched to their exact skin profile. The system scores every product in the catalog against the user's skin types, concerns, budget, and sensitivities. It then cross-checks the selected products against a set of known ingredient conflict rules before outputting the routine, so users never receive a recommendation that could harm them. Every product card in the output includes a explanation as to why that specific product was chosen for that specific user. Users can add products directly to cart from the routine view and check out on a single page.

Key Features: Some Key Features Cross Cleanse has is the Profile Builder- a 6-question quiz capturing skin type, top concerns, budget tier, ingredient sensitivities, experience level, and current actives with the user's skincare routine, the Recommendations- a scoring engine that ranks products by skin type match, concern overlap, budget fit, and sensitivity filters, the Flag Detector- flags 4 categories of dangerous ingredient combinations with severity levels and actionable tips, Routine generator- sequences products into the correct 6-step AM/PM application order with personalized reasoning per product, and the Product Catalog- users can shop directly from the catalog and find different products with the live filter system and add directly to their cart.

Technologies Used: The technologies used to build Cross Cleanse was in the Python, Flask, HTML, CSS, JS, NumPy, Pandas, CSS Keyframes, Kaggle, JSON and CSV datasets, Render, and Claude Code.

*Target Users: * The target users for Cross Cleanse is for all types of skincare users but don't know where to start and need a structured, safe routine built for them without buying a product that can irritate their skin.