mohit-karhana/PolicyGPT ? reverse-engineered prompt
Reverse engineered prompt
Build me an AI insurance policy assistant where I can create a policy, upload a PDF, and then ask questions in plain English like “Is diabetes covered?” or “What is the waiting period for pre existing diseases?”
I want the app to read the document in the background, turn the text into embeddings, and let me search it semantically or ask for a grounded answer that cites the exact page and section it came from. If the AI answer feature is not set up, the search part should still work.
Please make it a clean FastAPI backend with PostgreSQL and pgvector, background processing for PDF extraction, and Docker setup so I can run everything locally with one command. Add simple API routes for creating policies, uploading documents, checking processing status, searching, and asking questions. Also include basic validation for PDF uploads, consistent error responses, and tests so I can trust it works.
Are you gonna build this?
make sure you review the code using coderabbit