title: Find It Or Create It
emoji: 💻
colorFrom: green
colorTo: pink
sdk: docker
pinned: false
short_description: AI-powered Indian fashion search & custom designer booking.
👗 Find It or Create It ✂️ – AI Fashion & Custom Stitching Platform | 🌐 Live Demo
🧠 Overview
Find It or Create It is an AI-based Indian traditional fashion website that helps people either find similar dresses online or get their outfit custom stitched by a professional fashion designer. It mainly focuses on Indian traditional wear like sarees, lehengas, kurtas, sherwanis, and other cultural outfits. As we all know, in India, every region has its own fashion style — different types of embroidery, patterns, and designs. Because of this variety, it’s sometimes hard for normal people or even software to identify the exact dress style. To solve this, our website allows users to upload a photo or enter a text description of an outfit. Using AI image and text similarity models, the system searches and shows similar results from Amazon, Flipkart, and Google Images. If users still can’t find what they want, they can directly book a designer to create a similar outfit from scratch. This project is made only for portfolio and learning purposes, not for commercial use. It shows how AI and fashion design can work together in one platform.
🎯 Purpose
Indian traditional clothing is full of variety and culture. The same outfit can look completely different depending on which state it comes from — for example, a saree from Tamil Nadu and one from West Bengal have very different designs. Because of this diversity, finding the right outfit online is difficult. Our goal is to make this easy by: Using AI to understand what kind of dress or pattern the user is looking for. Showing similar dresses from online stores. Giving an option to design it personally if the user doesn’t find an exact match.
⚙️ How the Website Works
🔹 Step-by-Step Flow
1️⃣ Upload or Type
User can upload an image or type a text description (like “red embroidered lehenga”).
2️⃣ AI Search
Our system uses SigLIP + FAISS model to understand the image or text and find similar fashion outfits.
3️⃣ Show Results
If similar dresses are found, links to Amazon, Flipkart, and Google Images are shown. Users can open these links directly and check the dresses online.
4️⃣ Custom Stitching
If users are not happy with the results, they can go for “Customize & Stitch.” Here, they enter their name, contact, and measurements. Then the system automatically assigns a fashion designer to handle their order.
5️⃣ Designer Work
Once assigned, the designer receives a message with all customer details and starts working on the design. After stitching is done, the outfit is sent to the customer.
🤖 Automatic Designer Assignment
Whenever a user books a designer, our system automatically sends the details to the next available designer.
This saves time and avoids manual coordination.
Example Message 📢 New Task Assignment 📢
Hello Charlotte Davis,
Assigned Designer: Charlotte Davis (Employee ID: DES104)
A new customer request has been assigned to you automatically by our system.
--- 👤 Customer Details ---
Name: Hxxxx
Email: hxxxxi@gmail.com
Phone: 785xxxx21
Location: Hyderabad
--- 📐 Measurements ---
Chest: xx
Waist: xx
Shoulder: xx
Sleeve: xx
Selected Size: XS
--- 🛍️ Product Details ---
Search Query: Red Embroidered Pothys Poly Chiffon Lehenga Half Saree - Ladies' Lehenga for Special Occasions
--- 🆔 Assignment Info ---
Task generated automatically by system bot.
If you have any queries, please reach out to Krish (Team Lead).
⚡ Please proceed with this task at your earliest convenience.
Thank you! 🙌
🗺️ Roadmap (Website Flow)
🩵 Find It Path
👩💻 Upload Image / Text → 🤖 AI Detects → 🔍 Similar Outfits Found → 🛍️ Redirect to Amazon / Flipkart / Google
💜 Create It Path
👩💻 Upload Image / Text → 🤖 No Matches → ✂️ Choose “Customize & Stitch” → 🧵 Book Designer → 👗 Get Custom Outfit
🖼️ (A cartoon-style roadmap image can be added here to visually show both paths.)
🧮 Dataset Details
Used 10,000 sample data points (images and text).
To improve accuracy, more outfit data can be added.
The goal of this dataset is to show the working concept, not commercial accuracy.
🧰 Tech Stack
Layer Tools / Frameworks
Frontend HTML, CSS, JavaScript
Backend Flask (Python)
AI / Search FAISS, SigLIP
Image Optimization Pillow (PIL), Regex (for SVG minify)
Deployment Hugging Face Spaces
Data Storage JSON-based dataset
🧵 Features
This website brings together AI, fashion design, and usability in one place. Here’s what it currently supports:
✅ Upload an image or type text to search for similar outfits
✅ View matching products from Amazon, Flipkart, and Google
✅ Book a designer for custom stitching if no match is found
✅ Enter manual or automatic body measurements
✅ Automatic assignment of designer based on workload
✅ Optimized images (WebP, minified SVG) for faster loading
✅ JSON-based output storage for clean backend data management
✅ Fully functional and deployed on Hugging Face Spaces (Flask + Python)
✅ End-to-end tested workflow — from image upload to designer task assignment
🧑🎨 For Recruiters
This project is not a commercial website — it is a portfolio project created to demonstrate strong technical and creative skills in AI, web development, and fashion workflow automation.
It shows how AI-powered search can be connected with real-world fashion customization to build an intelligent and user-friendly experience.
🎯 Key Achievements
✅ Built an end-to-end multimodal search engine (text + image) completely from scratch
✅ Integrated Google SigLIP (Vision Transformer) model for accurate image–text alignment
✅ Implemented FAISS indexing for fast similarity search across 10,000+ embeddings
✅ Developed the complete backend pipeline: preprocessing → embedding → indexing → search → deployment
✅ Deployed securely using Docker and Hugging Face Spaces (Flask backend + AI model integration)
✅ Handled secret credentials safely through environment variables (no public exposure)
✅ Added automated designer assignment system using a JSON-based dataset
✅ Optimized all static assets (WebP + SVG minification) to improve performance on free-tier hosting
✅ Used JSON for structured data storage, ensuring easy debugging and lightweight processing
✅ Designed clear backend logs for customer requests and designer task reports
✅ Created optimized brand logo and visual assets to match website identity
✅ Achieved 3–5× faster loading speed through backend and image optimizations
✅ Maintained a clear roadmap and detailed documentation for future improvements
💡 Skills
🔹 AI-based image and text similarity search
🔹 Flask web application development and routing
🔹 Dataset preparation, embedding, and FAISS indexing
🔹 Frontend integration with Flask templates and routes
🔹 Fashion design workflow automation and user experience planning
🔹 Docker-based deployment and Hugging Face Space hosting
📧 Email: [mdeepa7879@gmail.com]
💼 LinkedIn: https://www.linkedin.com/in/deepa-m-66536a208/
license: This project is strictly for educational and portfolio purposes only, not for copying or commercial use, and anyone interested in viewing or using it must contact the creator.