---
title: "Overview"
description: "An intelligence platform that unifies your data and knowledge into an operational ontology."
icon: "diagram-project"
---

> **For AI agents:** the complete documentation index is at [llms.txt](/llms.txt). Append `.md` to any page URL for its markdown version.

The Aerovy Platform is an **intelligence platform**. It unifies the data and knowledge scattered
across your organization into a single **operational ontology** — a structured, living model of
your domain — and enables authoring **workflows** composed from that ontology's primitives. The
result: the institutional knowledge that used to live in people's heads and static documents
becomes **durable, executable tasks** that run on their own.

## What it unifies

The platform brings four kinds of input into one model:

| Input | What it is |
|-------|------------|
| **Hardware data** | Telemetry from connected devices: batteries, swap stations, chargers, meters, and vehicles. |
| **Third-party data** | Data pulled in from external systems through [Integrations](/console/integrations) — Geotab, Zubie, UtilityAPI, and others. |
| **Unstructured data** | Documents, specifications, and fileshares that describe how your domain actually works. |
| **Institutional knowledge** | The rules, expertise, and judgment your teams carry — the know-how behind day-to-day decisions. |

Everything is scoped to your organization.

## The operational ontology

These inputs are unified into an **ontology**: a single structured model that names and relates
the entities in your domain. It's built from a few reusable primitives:

- **[Thing Type Definitions](/platform/data-model#thing-type-definition)** describe the kinds of
  entities you operate — a battery, a charger, a product model.
- **[Metric Definitions](/platform/data-model#metric-definition)** describe the quantities they
  measure or derive — state of charge, tolerances, mean time between failures.
- **[Templates](/platform/templates)** map external and unstructured payloads onto those
  definitions so new data lands in the model consistently.
- **[Integrations](/console/integrations)** connect external systems to ingress data in and
  egress data and notifications out.
- **[Transactions](/console/transactions)** capture templatable records of business and
  operational events.
- **[Monitors](/console/monitors)** provide configurable, real-time observability over your
  metrics.
- **[Agents](/agentic/workflows)** are out-of-the-box models that execute specialized tasks —
  forecasting, anomaly detection, and more.

The ontology is **operational**, not a static graph with executable primitives that represent distinct units of work. 
The [Ontology Agent](/agentic/ontology-agent) builds durable context from your documents and data sources and
keeps it current as those sources change. See [Data model](/platform/data-model) for how the
primitives fit together.

## Workflows: knowledge as executable tasks

Once your domain is modeled, you **author [workflows](/agentic/workflows) composed from ontology
primitives**: a trigger plus a sequence of steps that query data, run agents, command hardware,
or call integrations. Each workflow encodes a real company task — alert response, a hardware
action, an integration handoff, scheduled planning — as a durable, executable version of it.

This is how institutional knowledge stops decaying: instead of living in a runbook or a single
expert, the decision logic is captured in a workflow that runs the same way every time, captures
performance data on each run, and improves as feedback arrives.

## The API

You interact with the platform through a single REST API, the **Aerovy Platform API**: one
HTTPS + JSON service for everything you do.

- **Write.** Register Things, Sites, Fleets, and Integrations; define your Thing Types and
  Metrics; and send telemetry.
- **Read.** Query a Thing's metric history, aggregate over time windows, and fetch latest
  snapshots.

The API is versioned in the URL path (`/v2/…`), and its base URL is environment-specific (for
example `https://spectra.dev.aerovy.com` in dev). See
[API fundamentals](/platform/api-fundamentals#base-urls-and-environments) for base URLs,
authentication, versioning, and errors.

## How data gets in

You don't always have to send data yourself:

- **Direct ingestion**: you call the telemetry endpoint with your own readings. This is the path
  covered in [Ingesting data](/platform/ingesting-data).
- **Integrations**: the platform pulls or receives data from a third-party system on your behalf
  (Geotab, Zubie, UtilityAPI, and others). You configure an **Integration** once, and the
  platform maps the external devices onto your Things automatically.

Once data is flowing, [Querying data](/platform/querying-data) covers reading it back out.