ChatbotLogo :LLM-Powered YouTube Video Q&A with Context-Based Answers from Transcripts by Using RAG

Niladri Ghosh1*,
1Ramakrishna Mission Vivekananda Educational and Research Institute (RKMVERI)   
Video Code

Key Features

  • Search Chats Instantly: keyword-based search over past conversations stored in ChromaDB.
  • YouTube Transcript Q&A: ask questions about a video's transcript using Retrieval-Augmented Generation.
  • LangChain-powered RAG: transcripts are split, embedded, and stored in ChromaDB before being queried via a LangChain chain.
  • Streamlit Interface: modern web UI for interacting with the chatbot.

Video

Getting Started

  1. Clone the repository from this link and install dependencies from requirements.txt.
  2. Follow the README.md for setup instructions.
  3. Run streamlit run app.py to launch the interface.
  4. Enter a search term or provide a YouTube link to start querying questions.