RAG
YouTube Video Q&A Assistant (LangChain RAG)
RAG assistant that summarizes YouTube videos and answers questions about their content.
An AI-powered assistant that fetches YouTube transcripts, generates structured summaries and answers questions about video content using Retrieval-Augmented Generation (RAG) with LangChain and OpenAI.
Highlights
- Fetches transcripts via
youtube-transcript-apiandyt-dlp, and stores video/channel metadata in SQLite with SQLAlchemy. - Chunks and embeds transcripts with OpenAI embeddings into a persistent ChromaDB vector store.
- Summarization (title, bullet summary, key takeaways) and scope-based Q&A across a single video or a whole channel.
- Three interfaces: CLI, FastAPI REST API and a Streamlit web GUI.
- Incremental processing skips already-processed videos.
- RAG
- LangChain
- ChromaDB
- FastAPI
- Streamlit
- YouTube