CodewithEvilxd/finguard ? reverse-engineered prompt

Build me a polished FinGuard AI app for spotting financial fraud and suspicious activity.

I want a public landing page, a demo sandbox, and a real dashboard where analysts can review transactions, alerts, investigations, and trends. The app should let users ingest transactions, check them for fraud and anomalies, show a clear risk score, and explain why something was flagged in simple terms. Please include an investigation view with account history, timeline, related entities, and analyst actions like confirm, clear, or escalate.

There should also be a grounded AI assistant that answers questions using approved internal policies and cites its sources, without making up transaction details. I want the whole thing split into a frontend, a core API, and an ML service, with a PostgreSQL database and a setup that works locally with Docker. If you need to, look up current docs online for the stack. Make it feel production ready, trustworthy, and easy to understand.