YouTube Q&A Assistant

A lightweight retrieval-augmented application that lets you extract transcripts from YouTube videos and ask questions directly against the video’s content. Built using LangChain, Groq LLaMA, FAISS, and the updated youtube-transcript-api.

📌 Overview

Long videos often take hours to watch and analyze. This tool helps you skip all of that by:

Fetching the video transcript

Splitting it into semantically meaningful chunks

Creating vector embeddings

Storing them in a FAISS index

Running retrieval-augmented generation (RAG) using LLaMA (Groq)

Answering your questions only from the video context

This makes it useful for research, learning, podcasts, interviews, and long-form discussions.

✨ Features

Up-to-date YouTube transcript fetching

Supports videos with English captions

Fast inference using Groq LLaMA 3.1

Chunked indexing for accurate retrieval

Clean separation of frontend and backend

Handles errors gracefully (no transcript, disabled captions, etc.)

🧱 Tech Stack

Python 3.x

LangChain (RAG pipeline)

Groq API (LLaMA 3.1 models)

FAISS (vector store)

youtube-transcript-api (updated .fetch() method)

Streamlit (frontend)