This project presents a Visual Rag System For Web Page Analysis designed to enhance the accuracy and relevance of AI-generated responses by grounding them in external data sources. Unlike traditional AI models that rely solely on pre-trained knowledge, this system dynamically retrieves relevant information from user-provided web content and generates context-aware answers.

The system allows users to input a webpage URL, from which content is extracted, processed, and divided into smaller chunks. These chunks are converted into vector embeddings using a local embedding model, enabling efficient similarity-based retrieval. When a user submits a query, the system retrieves the most relevant content and feeds it into a powerful language model via the Groq API to generate precise and contextually accurate responses.

Additionally, the project includes features such as quick summary generation, chat history, and multimodal support for images and text, making it more interactive and user-friendly. The system is implemented using Streamlit for the interface, Sentence Transformers for embeddings, and LLMs for response generation.

This approach improves reliability, reduces hallucination, and ensures that responses are grounded in real-time data, making it highly useful for applications like research assistance, study tools, and intelligent document analysis.