campusx-official/chatbot-in-langgraph ? reverse-engineered prompt

Build me a Python chatbot project that uses LangGraph and shows a few different chatbot demos in a simple Streamlit app.

I want one main chat experience, plus separate versions that demonstrate things like streaming responses, threaded conversations, using tools, saving chat data in a database, retrieval for answering from documents, and maybe an MCP style integration if that fits. Keep the UI clean and easy to use, with a chat box, message history, and a way to switch between the different examples.

Make it feel like a polished learning project, not a toy. The code should be organized so the backend logic is separate from the Streamlit front end, and it should be easy to run locally from one requirements file. If anything in the LangGraph or Streamlit setup needs current best practice, look up the latest docs online and use that.