talhaleet/LangGraph-AI-Chatbot ? reverse-engineered prompt

Build me a simple AI chatbot app in Python with a clean Streamlit interface. I want it to feel like a general assistant where I can chat normally, but it should also be able to use a few handy tools when needed, like web search, basic calculator help, and stock price lookup.

I also want to upload PDF files and ask questions about them, so the bot can answer using the document content instead of guessing. Please save chat history so conversations persist, using something lightweight like SQLite. A simple retrieval setup for the PDFs is fine, and the app should handle multiple uploaded documents in a practical way.

Use a LangGraph style workflow with tool calling, connect it to a Groq Llama model for responses, and use Google embeddings for the PDF search. Split things into a backend and a Streamlit frontend if that makes sense. Make sure it runs locally with environment variables for the API keys. If anything is unclear, look up the current docs online and make it work end to end.