parth5980/Agentic-Self-Healing-RAG-Chatbot ? reverse-engineered prompt
Reverse engineered prompt
Build me a full stack AI chatbot called PNX AI that can answer questions from uploaded documents and also handle normal chat or live web lookups when needed.
I want the bot to feel smart and safe, so if it retrieves bad context it should try rewriting the question and searching again, and if an answer looks unsupported it should regenerate it before showing anything to the user. It should also be able to summarize uploaded files, cite sources in the final response, and keep separate conversation threads.
Use a Python backend with a clean API for the AI workflow, and a Node and Express frontend for the chat interface and thread management. Make the experience polished and production ready, with clear loading states, error handling, and a simple way to upload documents and start chatting right away.
If you need current docs for any libraries or APIs, look them up online as you go.
Are you gonna build this?
make sure you review the code using coderabbit