A research tool where you can input a bunch of article URLs and ask questions. The tool retrieves answers based on those articles using Large Language Models (LLMs), embeddings, and vector search.
- Accepts multiple URLs of news articles
- Loads and extracts article content using LangChain’s UnstructuredURLLoader
- Constructs document embeddings using OpenAI
- Stores embeddings in FAISS for similarity search
- Accepts user questions and returns LLM-generated answers based on the most relevant articles
| Tool / Library | Purpose |
|---|---|
| Streamlit | Frontend interface |
| LangChain | Document loading and LLM integration |
| OpenAI API | Embeddings + Question Answering (LLMs) |
| FAISS | Fast Approximate Nearest Neighbor Search |
| dotenv | Secure handling of API keys |
| Python | Programming language used throughout |
git clone https://github.com/ShwetaPardhi0/LLM-News-Research-Tool.git
cd LLM-News-Research-Tool