🎧 YouTube RAG Assistant

This project is a Retrieval-Augmented Generation (RAG) pipeline built using Streamlit, Whisper, and Hugging Face Transformers.
It allows you to:

• Input any YouTube video link 🎥
• Automatically download and transcribe the audio using OpenAI Whisper
• Use sentence embeddings to create contextual knowledge chunks
• Ask natural language questions about the video content and get AI-generated answers 💬

🚀 Features

✅ Download YouTube video audio
✅ Transcribe audio → text using Whisper
✅ Create embeddings with Sentence Transformers
✅ Retrieve the most relevant context for your question
✅ Generate accurate, summarized answers using Flan-T5
✅ Interactive Streamlit UI

🧱 Project Structure

youtube-rag-assistant/
│
├── rag.py                 # Core RAG pipeline (transcription, embedding, QA)
├── app.py                 # Streamlit UI for interaction
├── requirements.txt       # Dependencies
└── README.md              # Project overview and usage guide

⚙️ Setup Instructions

1️⃣ Clone the Repository

git clone https://github.com/Japan1907/youtubevideo-rag.git

2️⃣ Create a Virtual Environment

python -m venv .venv Activate it

3️⃣ Install Dependencies

pip install -r requirements.txt

4️⃣ Run the Application

streamlit run app.py