Intelligent YouTube Video Summarizer and QA System
Published:
Developed a comprehensive application that processes YouTube videos to make their content searchable and conversational. The system automates the downloading of videos, transcribes the audio, and uses LLMs to generate concise summaries. Key features include a semantic search capability powered by a FAISS vector database and an interactive chat interface for asking questions directly about the video’s content, facilitated by LangChain. The system is accessible via a web UI, a RESTful API, and a command-line interface.
Tech Stack: FastAPI, Streamlit, LangChain, FAISS, SQLAlchemy, Redis, Groq API, Nvidia Inference API.