This project is a URL Summarizer and Q&A system that processes web articles from given URLs and answers user questions using retrieval-augmented generation (RAG).
The system uses LangChain for retrieval, Hugging Face's google/flan-t5-base for answer generation, and FAISS for efficient document indexing.
The graphical user interface (GUI) is built using Streamlit.
- URL Processing for extracting article content
- Question Answering with natural language queries
- Document Retrieval using FAISS vector index
- Answer Generation with
google/flan-t5-base - Clean GUI displaying questions and answers
- Local Model Execution for privacy and speed
- URL Input: Users provide up to three article URLs.
- Content Extraction: The system loads and processes the URLs.
- Text Splitting: Articles are split into chunks for indexing.
- Vector Indexing: FAISS creates a vector index of document embeddings.
- Question Processing: Users ask a question about the articles.
- Retrieval and Generation: Relevant chunks are retrieved, and answers are generated.
- GUI Display: The question and answer are shown in the Streamlit interface.
Ensure you have Python installed (version 3.9+ recommended). Install dependencies using:
pip install -r requirements.txtstreamlit run main.py
