TERRANOETIS

Getting started

Platform v3.0.0 Docs v3.0.0

This is the narrative companion to Getting started (site page), which carries the verified commands, measured outputs and feature-availability tables. Read this file for context; trust the page for numbers.

What Terranoetis is

Terranoetis is a real-time geospatial intelligence platform. A React 19 + CesiumJS client renders a WebGL globe; an Express/TypeScript API serves live data from public providers, a catalog of 150 literature-cited analytical equations, and seven hazard-physics simulation kernels; the same API hosts an AI chat pipeline with dual-process (System 1 / System 2) cognition and continuous monitoring engines. Everything runs on your machine: the analytical engine and three simulation kernels need no external credentials at all.

Mental model

Three facts orient everything else:

  1. The API is the product boundary. The globe UI and any script you write hit the same /api endpoints. If you can curl it, you can automate it.
  2. Simulations are kernels, not cloud services. Each hazard is a Python file in kaggle-kernels/. Three of them (earthquake, wildfire, hurricane) are executed locally by the server with python3; four (flood, tsunami, volcano, landslide) are dispatched to Kaggle as an execution venue — the code itself is pure NumPy, run on CPU instances except flood-sim, whose kernel metadata requests Kaggle's GPU accelerator.
  3. Nothing is fabricated silently. Inputs that are required but missing raise at the client boundary; analytical tools that cannot obtain real data return NaN with an explanation. The docs mirror this: every claim on the site carries a file:line citation, and unexecuted claims are labelled.

First hour

  1. Install and run: the commands and the measured health payload are on the site page.
  2. Run one simulation end-to-end: a curl earthquake job completes in well under a second with no keys configured — the page shows the transcript.
  3. Read one capability page top to bottom (start with earthquake) to see what a fully-cited page contains: equations, parameter contract, outputs, validity limits, measured runs, reproduction.
  4. When you need raw endpoints, use the API reference and /api/openapi.json; when you need the simulation wire format, use the simulation job API.

Where each document lives

Markdown in this repository is the source of truth for prose; the HTML site (generated from scripts/docs/) carries structured, verified data and navigation. If you edit prose, edit Markdown; if you edit a parameter table or an equation, edit the page spec under scripts/docs/ and rebuild (node scripts/docs/build.mjs), then pass the quality gate (node scripts/docs/quality-gate.mjs).