🔍 HTML Search SPA

A full-stack semantic search tool for crawling, chunking, and vectorizing HTML content from any public website URL.

🎥 Demo

🌐 Features

• Accepts any public website URL.
• Parses and chunks HTML content.
• Indexes with Weaviate vector database.
• Lets you search semantically across website content.
• Shows match percentage, context preview, and raw HTML.

---

🛠️ Tech Stack

• Frontend: Next.js
• Backend: Python (Flask)
• Vector DB: Weaviate
• NLP Embedding: HuggingFace SentenceTransformer (all-MiniLM-L6-v2)

---

🚀 Running the Project

1. Backend

cd backend
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
python app.py

2. Frontend

cd frontend
npm install
npm run dev

3. Vector DB

docker-compose up -d

📁 Folder Structure

html-search-spa/
├── backend/
│ ├── app.py
│ ├── utils/
│ ├── requirements.txt
├── frontend/
│ ├── public/
│ ├── src/
│ │ ├── App.jsx
│ │ ├── components/
│ ├── tailwind.config.js
│ ├── package.json
├── weaviate/
│ ├── docker-compose.yml
├── README.md
├── .gitignore

To Push to GitHub

git init
git add .
git commit -m "Initial commit: HTML Search SPA"
git remote add origin https://github.com/your-username/your-repo.git
git push -u origin main

📃 License

MIT License © 2025 Sayan Debnath