LLM URL Question Answering (Colab Project)

This project is an LLM-based question answering system that reads content from multiple URLs and answers user questions using only the extracted webpage content.

The implementation is done in a Google Colab notebook.

What this notebook does

• Loads webpage content from given URLs
• Cleans and splits the text into chunks
• Uses LangChain with HuggingFace models
• Answers questions using strictly retrieved context
• Helps reduce hallucinations

How to run

1. Open the notebook in Google Colab
2. Install required libraries
3. Run cells from top to bottom
4. Enter URLs and a question

Technologies used

• Python
• Google Colab
• LangChain
• HuggingFace Transformers

Dependencies

This project lists only the primary dependencies required to run the notebook. Some libraries (e.g., tokenizers, sentencepiece, unstructured sub-dependencies) are installed automatically as transitive dependencies.