Shreeyans2305/SchnellXAI — reverse-engineered prompt

Build me a full stack logistics control app called SchnellXAI. I want a dashboard where an operator can watch shipments and hubs in real time, see delays, congestion, carrier issues, ETA drift, and other disruption signals, and have an AI pipeline review those signals every 25 seconds.

The app should feel like an operations tool, not just a chatbot. Show shipment health, active anomalies, decisions being made, a live event feed, charts, and a map view. The AI flow should follow five stages, observe, reason, decide, execute, and learn. Low risk actions like notifying carriers or rerouting a single shipment can happen automatically, but bigger reroutes that affect several shipments should go into a human approval queue with the proposed route and updated ETA clearly shown.

Please include short term memory for recent events, long term learning in a local database, action logs, approvals, and disruption history. Use a local model through Ollama for the AI side. If real APIs are not available, seed it with realistic demo logistics data and disruption simulation so the whole experience works end to end. Look up current docs online if you need to.