login
Visual finder
Product finder
Try on
low-lvl: cloth-segmentation example
low-lvl: whisper to query
fashIA examples

ECHOLOOK 👗✨

The AI-powered fitting room in your pocket

🚀 Inspiration

The online fashion-shopping experience still lacks the closeness and spontaneity we feel when trying on clothes in a store. We wanted to bring the fitting room to your phone: just snap a photo and speak to discover Inditex garments that match you, see them on your body in seconds.

🛠️ What it does

Feature	Flow
🔐 Sign-up / Log-in	Secure JWT auth
📸 Smart photo	Capture / upload → U²-Net mini segments clothes and spots centroids
🔍 Visual search	Click a centroid → Inditex Visual Search API serves similar garments (optimized with segmentation!)
🎙️ Voice search	Record any language → Whisper transcribes → Gemini turns text into a structured query for the Product Search API
🪄 Virtual try-on	One tap → Fashn.ai blends suggested garments onto your own photo
🗂️ Personal history	Every photo, mask & suggestion is auto-saved for later

🧰 How we built it

Frontend: Vue + Quasar Backend: Node.js & Express, S3 integration Database: MongoDB AI & micro-services cloth-segmentation-mini (U²-Net) whisper-query-parser (Whisper-small + Gemini Pro) served with FastAPI inside Docker Fashn.ai API for virtual try-on Inditex Visual Search and Product Search APIs

😅 Challenges we ran into

Wiring up multiple external dependencies and keeping the whole app in sync
Picking the right models for speech-to-text, the language model, and the generative model used for virtual try-on

🏆 Accomplishments that we're proud of

First end-to-end demo: from photo + audio to virtual fitting room A segmentation model that boosts the precision of Visual Search Thought-through UI/UX design Modular architecture: each AI service can run as an independent API

📚What we learned

Generative AI is amazingly powerful, but the magic only happens when you wrap it in a frictionless UX. Shoppers value being able to see clothes on their own body far more than just browsing images.