🐍 Python Video Q&A Chatbot

An AI-powered tutor for Python YouTube videos — ask, watch, learn, and code.

An interactive chatbot that lets you ask natural language questions and searches for answers in youtube videos — powered by LangChain, OpenAI, and ChromaDB. Specialised in Youtube videos for learning python programming.

---

🚀 Features

• 🔗 Paste a YouTube video URL about Python coding and embed it directly in the app
• 🧠 Ask questions about the content using natural language
• 📖 Vector search over transcript chunks with timestamp + chapter metadata
• 📺 Plays part of the video that answers the question
• 📟 Additional explanations and coding challenge
• 🤖 Powered by Llama3 8B + LangChain RetrievalQA

---

📊 View Project Presentation Slides

---

🛠️ Tech Stack

Component	Description
Streamlit	Frontend UI for chat, video player, and code interaction
LangChain	Retrieval-Augmented Generation (RAG) orchestration
LangSmith	Tracing and debugging of LLM chains and prompts
ChatGroq (LLaMA3-8B)	LLM used for answering questions and generating challenges
OpenAIEmbeddings	Converts transcript chunks into vector representations
ChromaDB	Local vector database for storing per-video embeddings
pytubefix	Downloads captions and extracts video metadata
GPT-4 (optional)	Evaluates the quality of LLaMA3 responses post-hoc

---

📦 Setup Instructions

1. Clone the repo

   git clone https://github.com/KJanzon/youtube-qa-chatbot.git
   cd youtube-qa-chatbot

2. Set up virtual environment

   python -m venv venv
   source venv/bin/activate  # or .\venv\Scripts\activate on Windows

3. Install dependencies

   pip install -r requirements.txt

4. Add your API key Create a .env file with:

   OPENAI_API_KEY=your_openai_key_here
   LANGCHAIN_API_KEY
   HUGGINGFACEHUB_API_TOKEN
   GROQ_API_KEY
   

5. Run the app

   streamlit run interfaces/streamlit_chat.py

---

📁 Folder Structure

├── app/                  # Video processing + transcript embedding
├── data/                 # Downloaded caption files (.srt)
├── interfaces/           # Streamlit front-end
├── utils/                # Helpers (e.g., clean_srt, time utils, chapter ranker)
├── vectorstore/          # ChromaDB persistent store (per-video)
├── .env                  # API key config (excluded from Git)

---

📸 Screenshot

🧠 TODO / Roadmap

• Multi-video querying (cross-video RAG)
• Reference official python tutorial to ensure correct answers (https://docs.python.org/3/tutorial/index.html)