laxmimerit/Agentic-RAG-with-LangGraph-and-Ollama — reverse-engineered prompt

Reverse engineered prompt

GitHub

Build me a local AI assistant for chatting with my own documents, using Ollama so my data stays on my machine. I want a simple Streamlit screen where I can upload PDFs or text files, ask questions, and get clear answers with the sources it used.

Please make it work like an agent, not just a basic chatbot. It should decide when to search the documents, when to answer directly, and when to say it doesn’t know. Add conversation memory so it remembers the current chat, and include a simple way to save longer term notes if that fits. If possible, include a small MySQL example where the assistant can answer questions from a database too.

Use LangChain and LangGraph for the workflow, FAISS for document search, and Ollama models like Llama or Qwen. Include setup instructions, required installs, and a few sample files so I can run it locally. Look up current docs online if you need to.

Want more depth? Deep Reverse