WebChat+ - Chat with Websites from URL

WebChat+ is an AI-powered web assistant that I built that allows anyone to interact with and extract meaningful insights from website and it's content. Built using Streamlit, LangChain, and Huggingface, WebChat+ leverages state-of-the-art language model from Mistral (Mistral-7B-Instruct-v0.2) and embeddings to facilitate intelligent conversations based on retrieved website data.

🚀 Features

• Website Content Processing: Extracts and tokenizes website content for efficient retrieval.
• Conversational AI: Uses a retrieval-augmented generation (RAG) model to answer user queries based on website data.
• Hugging Face Integration: Utilizes models from Hugging Face Hub for intelligent response generation.
• Vector Store for Retrieval: Employs ChromaDB for semantic search and fast document retrieval.
• Streamlit UI: Provides a user-friendly and interactive chat experience.
• Real-Time API Status Check: Displays system and API connection status.
• Customizable Settings: Supports dynamic website input and personalized interaction.

🛠️ Installation & Setup

Prerequisites

Python 3.8+
pip
A Hugging Face API token 

Clone the Repository

git clone https://github.com/abu14/chat-with-website-langchain
cd WebChat+

Requirements

pip install -r requirements.txt

Set Up Environment Variables

Create a .env file in the root directory and add your Hugging Face API token:

HUGGINGFACEHUB_API_TOKEN=your_api_token_here

Tech Stack

Frontend: Streamlit
Language Models: Hugging Face Hub (e.g., Mistral-7B-Instruct-v0.2)
Embeddings: Hugging Face Embeddings (e.g., sentence-transformers/all-MiniLM-L6-v2)
Vector Store: Chroma
Text Processing: LangChain, BeautifulSoup, RecursiveCharacterTextSplitter
Environment Management: dotenv

🔧 Usage

Run the Streamlit app

Once the application starts, follow these steps:

1. Enter a website URL in the sidebar.
2. Wait for content processing to complete.
3. Ask questions related to the website content.

📌 How It Works

• Content Loading: Extracts text from the given website URL.
• Text Splitting: Breaks the content into manageable chunks for better retrieval.
• Embedding Generation: Converts text into vector embeddings using Hugging Face models.
• Vector Storage & Retrieval: Stores the embeddings in ChromaDB for quick semantic search.
• Conversational RAG Model: Uses LangChain to retrieve relevant content and generate responses.
• Response Processing: Filters and cleans responses before displaying them in chat

📷 Screenshot

📞 Contact

Developed by Abenezer Tesfaye

• GitHub: abu14
• Email: tesfayeabenezer64@gmail.com