Skip to main content

Getting Started

This guide provides the primary onboarding path and technical prerequisites for integrating with the RiverGen AI Ecosystem.

Overview: The Intelligence Layer

The RiverGen AI runtime is segmented into three distinct components. Instead of general algorithms, you have ready-to-deploy, high-level interfaces tailored for practical software integration.

  • Model Studio: Your central hub for dataset injection, auto-training (AutoML), and tracking machine learning experiments without managing complex Python codebases.
  • Prompt Studio Agent (PSA): A secure chat interface that converts natural language into safe database queries and API calls. It serves as an autonomous SQL and Data Assistant.
  • Decision Intelligence (DIA): An active monitoring engine that evaluates incoming data against your trained models in real-time to execute automated corrections or alerts.

Prerequisites

Before interacting with the AI ecosystem, ensure you have the following prerequisites configured across your local workspace or staging environment.

Database Connections

Required: You must have an active PostgreSQL connection. Model Studio stores metadata and DIA records inference logs in structured persistence tables. Ensure your local .env contains a valid DATABASE_URL.

API Token Configuration

If interacting with the hosted version of RiverGen AI, navigate to Settings > API Keys inside the dashboard to generate a secure Bearer token for automated endpoint testing and integration.

Your First AI Workflow

The following four-step process describes the end-to-end flow of connecting datastores, training an AI model, and actively querying the environment.

1. Upload and Prepare via Model Studio

Connect a CSV dataset or stream telemetry data into your instance. The Model Studio intelligently handles missing data, automatically creates feature sets, and prepares the core algorithm structures.

Learn more about Model Studio Overview

2. Train and Register

Execute the training pipeline. RiverGen identifies the best models (e.g., Random Forest or XGBoost) to maximize accuracy. Once initialized, the model is pushed into the unified platform "Registry".

Learn more about Registry Management

3. Interact via Natural Language

Switch to the Prompt Studio Agent to interact with your data. Instead of writing complex SQL, you can ask questions like "Show me the anomaly rates across our recent models over the last 7 days" to retrieve secure insights.

Explore the Prompt Agent Architecture

4. Deploy to Decision Engine

Take the finalized model from the registry and attach it to standard ingestion pipelines via Decision Intelligence (DIA). DIA will begin monitoring platform actions entirely autonomously.

Deep dive into DIA

Next Steps

Ready for deep technical integrations? Select a module to explore its respective implementation workflows, architecture paradigms, and standard engineering practices.


  • (c) 2026 RiverGen. Confidential - For Internal Documentation and Technical Integration.*