Welcome to the LLM-Based Web Search project! This project leverages a Language Model (LLM) to perform real-time search and retrieval of information from the web. It showcases the power of combining LLMs with web scraping and retrieval-augmented generation (RAG) techniques to provide accurate and contextually relevant information.
You can access the demo of the Streamlit app and experience the capabilities of LLM-based web search firsthand.
🎉 Try the Demo on Hugging Face Spaces
The backend for this app is also hosted on Hugging Face Spaces. You can explore the API that powers the LLM-based search and see how the magic happens behind the scenes.
This project utilizes advanced generative models to interpret search queries and fetch relevant data from the web. The process involves:
- Search & Retrieval: The app scrapes web content based on user queries.
- Content Processing: Extracted content is then processed and filtered.
- LLM Generation: The cleaned and concatenated content is fed into an LLM, which generates insightful responses.
The frontend of this project is built with Streamlit, providing a simple and interactive interface. Users can input queries and receive detailed responses in real-time, powered by the LLM.
The backend is developed using Flask, serving as an API that handles the heavy lifting of web scraping, content processing, and interaction with the LLM.