README.md

---
title: Visual Image
emoji: ⚡
colorFrom: green
colorTo: blue
sdk: docker
pinned: false
---

# Visual Image Search

A powerful image search application using CLIP embeddings and Qdrant vector database. Search for images using natural language descriptions or upload similar images to find matches.

## Features

- 🔍 **Text-to-Image Search**: Search images using natural language descriptions
- 🖼️ **Image-to-Image Search**: Upload an image to find similar images
- 🌐 **URL Search**: Search using image URLs
- 📁 **Folder Management**: Index and manage multiple image folders
- ⚡ **Real-time Updates**: WebSocket support for live indexing progress
- 🎯 **Advanced AI**: Powered by OpenAI's CLIP model for semantic understanding

## How to Use

1. **Add Folders**: Click "Add Folder" to index your image directories
2. **Search by Text**: Type descriptions like "a red car" or "sunset over mountains"
3. **Search by Image**: Upload an image to find similar ones in your collection
4. **Browse Results**: View thumbnails and original images

## Technology Stack

- **Backend**: FastAPI, Python
- **AI Model**: OpenAI CLIP for image-text embeddings
- **Vector Database**: Qdrant for similarity search
- **Image Processing**: PIL/Pillow
- **Frontend**: HTML, CSS, JavaScript with real-time WebSocket updates

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference