Run Preempt Analytics on your own machine.

Everything runs inside Docker — no Python, no ML knowledge needed. Just two quick installs, then some waiting while things download.

Model Retrains

live · updated hourly

CI Pipeline Runs

live · last run 41 days ago

You don’t need to know how to code. The whole journey is three steps:

1. Install Docker Desktop and Git (below) — two free apps, installed like any other program.
2. Get the code onto your computer — paste one command into a terminal (explained below).
3. Run the menu script and press a number — press 6 first to turn everything on, then any other number to try something.

That’s the whole thing — no Docker commands to memorize, just a couple of lines to paste and a number to press after that.

You need two things, both free, both one-time installs:

Docker Desktop — install it and make sure it’s running (the whale icon in your menu/task bar).
Git — used once, to get the project folder onto your computer.

Get the code onto your computer

Open a terminal and paste this:

git clone https://github.com/Preempt-Analytics-Demo/predictive-maintenance-demo.git
cd predictive-maintenance-demo

That folder is your project folder for everything below. Prefer to browse first? View the source on GitHub →

GitHub is a website where code is stored and shared — think of it like a shared drive, but for software projects.

Cloning just means downloading a full copy of that project onto your own computer — the git clone command above does exactly that.

A terminal (or "command line") is a text-based way of giving your computer instructions. Paste the two lines above in, one at a time, and press Enter after each.

The easy way — a menu, no commands to remember

Once you have the project folder on your computer (see above), open a terminal in that folder and run one script below. It replaces every Docker command with a numbered menu.

# Mac / Linux
./preempt.sh

# Windows
.\preempt.ps1

Press 6 first — that turns on the background services (once per session is enough). After that, press 1 or 2 to simulate sensor readings, 3 to open the drift report, or 4/5 to watch things happen live.

Prefer to type it? Three commands

This replaces everything above — get the code and start the stack in one go, no menu needed.

git clone https://github.com/Preempt-Analytics-Demo/predictive-maintenance-demo.git
cd predictive-maintenance-demo
docker compose up -d

From there you can generate sensor readings and watch drift detection trigger an automatic retraining run — the full walkthrough lives in the README.

Heads up: retraining runs on a shared demo pipeline (about 10 runs/hour for everyone combined). If that limit’s been hit, your data still uploads — this run’s retraining is skipped rather than queued, and the limit frees up again on a rolling basis. Nothing is broken.

Full instructions on GitHub →