Youtube-rag

This repository is a simple RAG system built to answer questions from a youtube video.

How to setup:

• Simply clone the repo
• Install the required libraries
• Create a .env file and add these two environment variables:

  OPENAI_API_KEY=""
  PINECONE_API_KEY=""
  

  Copy your keys in it.
• Setup free pinecone account. Create a simple project and in that project create an index. This name will be provided at the input.

Running app:

python rag.py --url {video_url} --index {index-name} --model {model name default: gpt-3.5-turbo} --transcript {path to create transcript file}

Below is a demo of output: