Rakaa503/interaction-human-with-ai ? reverse-engineered prompt

Build me an AI interaction app called AVIGO that can receive a user message, figure out what the user means, and then decide how to respond.

I want the main backend in Go, with an API that handles conversations and interactions, saves everything to PostgreSQL, and acts like the orchestrator for the whole flow. The app should call a separate Python service for the machine learning part, where the message gets classified into intent, emotion, and topic. If the ML service can’t decide confidently, use a simple rule based fallback.

Please include a basic conversation context layer so the app can remember recent messages and a decision layer that can route the result into actions like greeting, answering a question, solving a problem, handling a request, general chat, or asking for clarification.

Also set up the training and prediction side in Python, with a small sample dataset and a model file that can be retrained. Make it easy to run locally, with health checks for both services and clear setup instructions. If you need current docs, look them up online.