ClosetMuse

ClosetMuse - AI Personal Stylist powered by Google Gemini 3
Atelier Exclusives - Bespoke Outfit Curation
Style Audit - Strategic Gaps
Color Analysis - Find My Palette
Atelier Muse - The Master Wardrobe

🌟 Introduction

ClosetMuse was born from a simple question: Why does styling yourself still feel overwhelming when your wardrobe is full?

Most people don’t lack clothes - they lack visibility, guidance, and aesthetic clarity. We wanted to build a system that feels like a personal stylist, creative director, color expert, and beauty consultant, all rolled into one beautifully crafted experience.

That vision evolved into ClosetMuse, an AI-powered stylist built around deep multimodal understanding using Google Gemini 3.

🌱 Inspiration

While exploring fashion tech, we noticed a recurring pattern:

Wardrobe apps rely too much on manual tagging.
Outfit generators ignore texture, silhouette, and composition.
Color analysis tools feel outdated and generic.
No single tool bridges fashion, beauty, and personal color harmony.

We wanted to create something more human, more intuitive, and more intelligent — something that felt like stepping into a curated atelier built just for you.

🧠 What We Built

ClosetMuse consists of four interconnected systems, each designed to mirror real styling workflows.

1. 🧵 Atelier Muse - The Master Wardrobe

A beautifully visualized digital closet that automatically transforms raw photos into clean, categorized wardrobe sections.

Key features:

Ingests outfits from images and videos
Automatic background removal
Smart categorization (Tops, Outerwear, Dresses, Shoes, etc.) and de-duplication of clothes
Clean virtual clothing rails for aesthetic browsing
Designed to feel like your personal boutique rack

This lays the foundation for AI-driven styling.

2. ✨ Atelier Exclusives - Bespoke Outfit Curation

Powered by Gemini 3’s multimodal reasoning, this engine understands more than color or tags - it sees fabric, drape, sheen, texture, and silhouette.

What it does:

Generates curated looks (“Corporate Power Look”, “Winter weekend brunch”, etc.)
Balances colors, proportions, and style logic
Writes stylist-style composition notes
Supports virtual try-on using your digital twin

This makes everyday styling feel editorial, inventive, and personalized.

3. ⚡ Style Audit - Strategic Gaps

Most people don’t need more shopping, they need smarter shopping.

ClosetMuse identifies Multiplier Pieces: items that unlock the highest number of new outfits.

You get targeted suggestions like:

White Tailored Button-Down
Black Pointed-Toe Boots
Cream Structured Trench Coat

The Google shopping links are also provided for the suggested items so that the user can shop these items or save them in their wishlist for later purchase. This feature transforms wardrobe planning into a strategic, data-backed process.

4. 🎨 Color Analysis - Find My Palette

This system elevates ClosetMuse beyond wardrobe guidance into full aesthetic alignment.

Using Seasonal Analysis, Gemini evaluates:

Skin undertones
Eye clarity
Natural contrast
Temperature, depth, and chroma

It determines your season (e.g., Winter, Summer, Autumn, Spring) and builds a deeply personalized color profile.

Color Analysis includes three pillars:

Color Harmony: Determines which colors visually enhance your natural features, and which tones dull or flatten them.

Find My Palette: Generates a Power Palette with season-specific tones (e.g., bold jewel colors for Winter, soft cool pastels for Summer).

Makeup & Hair Recommendations: Based on your palette, the app suggests

Foundation undertone
Lipstick shade family
Bronzer & blush tones
Hair color options (e.g., Jet Black, Deep Espresso, Cool Burgundy)

The suggested items come along with the Google shopping link for direct purchase through the app. This aligns fashion, beauty, and personal identity into one cohesive aesthetic system.

🔧 How We Built It

Google Gemini 3 (The Core)

Google Gemini API: ClosetMuse is built on Google AI Studio, entirely around the advanced multimodal capabilities of Google Gemini 3, transforming it from a standard inventory app into an intelligent visual engine.
- We leverage Gemini 3 Pro Preview for high-level reasoning tasks that require nuanced visual understanding. It powers our Deep Color Analysis, interpreting subtle skin undertones to recommend seasonal palettes, and our Smart Garment Detection, which identifies and categorizes multiple items from a single cluttered photo using structured JSON outputs.
- For the visual experience, Gemini 3 Pro Image Preview is central. It drives the "Atelier Studio-izer", which edits raw user photos into clean, background-free assets (Ghost Mannequin effect), and the Virtual Try-On, which uses complex image editing to visualize recommended outfits on the user's digital twin while preserving their identity.
- Finally, Gemini 3 Flash Preview ensures our Outfit Curator and Style Audit features provide instant, context-aware styling advice, making the application feel responsive and alive.

Frontend

React 19: Component-based UI library.
TypeScript: Ensures type safety for complex data models (Clothing Items, User Profiles, Outfit Recommendations).
Tailwind CSS: Utility-first CSS framework used for the luxury aesthetic, responsive grid layouts, and glassmorphism effects.
Lucide React: Beautiful, consistent iconography used throughout the app.

Backend & Infrastructure (Serverless)

Firebase Authentication: Handles secure user sign-up, login, and session management.
Cloud Firestore: NoSQL database used to store user profiles, wardrobe item metadata, styled outfits, and wishlist data in real-time.
Firebase Storage: Scalable object storage for hosting user-uploaded wardrobe photos and videos, profile pictures, and generated try-on results.

Flow Diagram

ClosetMuse Flow Diagram

⚡ Challenges

Teaching the model to distinguish textures, drape, and sheen
Ensuring consistent wardrobe categorization despite low-quality inputs
Extracting multiple clothing items from uploaded photos and videos
Designing the wardrobe gap feature to suggest outfits that truly elevate the outfit

🎯 What We Learned

How to use multiple Gemini 3 models for different use-cases
How to integrate Firebase for backend in the Google AI Studio
How to ingest and store the photos and videos without degrading the quality
How to integrate Google Search with the suggested items for generating shopping links
How color science, wardrobe math, and deep reasoning can work together

🎉 Accomplishments that We’re Proud Of

Built a fully functional AI stylist end-to-end. From wardrobe digitization to outfit generation, style-gap analysis, and seasonal color analysis.
Integrated Gemini 3 multimodal capabilities to understand clothing texture, color, drape, fit, and aesthetic context—far beyond simple metadata tagging.
Designed a digital twin visualization flow for “try-on without trying,” enabling users to preview curated looks realistically.
Delivered a product that bridges fashion + engineering, turning complex styling logic into a simple, elegant experience.

These wins made the project feel like a real, consumer-ready app — not just a prototype.

🚀 What’s Next for ClosetMuse

AI-Powered Outfit Calendar: Plan your weekly outfits based on weather, mood, events, or travel itineraries.
Closet Health Score: A sustainability-inspired metric showing:
- how often you wear each item
- cost-per-wear
- underused pieces
- items worth donating/selling
Friend & Partner Wardrobe Sync: Allow users to share closets and create coordinated outfits.
ClosetMuse Pro A premium tier offering:
- trend forecasts
- exclusive “Atelier Exclusive” lookbooks
- advanced beauty palette matching

The long-term goal is to transform ClosetMuse into a holistic personal style ecosystem, from wardrobe to makeup to shopping, and eventually into a platform that learns your identity and evolves with you.