$ docker pull aiopinc/koios:latest

Industrial Intelligence, Reimagined

Ai-OPs deploys real-time AI at the edge — on your existing control systems, with full data sovereignty.

v1.2.0GA · September 16, 2026

Koios 1.2

One workspace for building components, folders to organize them, and isolated package stacks so a component can run against a library version Koios does not ship.

Read the release notes

workspace

Components build in one view

An environment rail, the canvas, and an inspector in one three-pane workspace. Saving a running environment applies in place, so component timers and buffers survive the change.

stacks

Isolated package stacks

Attach an environment to a named set of wheels you upload. Its components run against exactly those versions, checked before anything uses them.

models

Differently-sampled inputs

A model file can carve one input vector into named inputs, each with its own window length, sample rate and interpolation. A per-second flow and an hourly temperature feed the same model.

devices

Serial and GPIO hardware

Reach serial, GPIO, I2C and SPI devices on the host without a world-writable device node or a privileged container. OPC-UA moves onto its client library's first full release.

security

Audit log and certificates

An append-only record of who did what, with its own retention and export, plus one place to install the TLS certificate browsers see when they connect.

trends

Exports that stream

A trend export downloads as it is produced rather than being built on the server and left behind, so larger ranges finish instead of being lost.

Everything you need at the edge

From data collection to AI inference to real-time visualization — purpose-built for industrial operations.

Real-time AI Inference — capability preview

Real-time AI Inference

Deploy in minutes

Pull the image, configure the service, and you're running — full platform deployment in under five minutes.

koios — deploy
Koios dashboard
Real-time Dashboards
Model Management
Historical Trending
Device Configuration
30,000+
hours of production runtime
600%
average return on investment
<1%
added management overhead

Trusted by industry leaders

The Ai-controller helped the hydrotreater process unit operate with more stability producing better heat distribution through the exchangers, less flow valve cycling and longer run times.

Bobby Newsome
Manager, Transmix — Sunoco LP

Koios gave us real-time visibility into our process optimization that we never had before. The deployment was straightforward and the ROI was measurable within the first quarter.

Operations Lead
Refinery Operations

Running AI models on-prem was a non-negotiable requirement. Koios was the only platform that delivered edge inference with the industrial protocol support we needed.

Controls Engineer
Power Generation

Ready to Transform Your Operations?

See how Ai-OPs can deliver real-time AI optimization to your facility.

Contact Us Today