Introduction

Artemis is a no-code AI platform for ROS 2 robots. Train, simulate, and deploy machine learning models to your robot fleet — without writing a single line of ML code.

What is Artemis?

Artemis sits between your robot hardware and your team. You upload sensor data, Artemis trains a model, and the result gets pushed over-the-air to your ROS 2 nodes — all from a browser dashboard. No MLOps background required.

Whether you're prototyping a single robot arm or managing a fleet of 50 autonomous vehicles, Artemis scales to fit your workflow.

Key Features

• No-code training — Upload your dataset and Artemis automatically selects the right model architecture, trains it, and packages it for ROS 2 deployment.
• ROS 2 native — Artemis integrates directly with ROS 2 topics and services. Deploy models as standard ROS 2 nodes — no custom bridges needed.
• Hardware-in-the-loop simulation — Test your trained model in a physics-accurate simulation before pushing to live hardware, preventing costly crashes.
• OTA model updates — Push new model versions to your fleet over-the-air with zero downtime. Artemis handles rollbacks automatically if performance drops.
• Real-time monitoring — Track inference latency, resource utilization, and behavioral anomalies across your entire fleet from one dashboard.

How It Works

Artemis follows a four-stage pipeline that maps directly to the way robotics teams work:

Prerequisites

Artemis is designed to work with any ROS 2-compatible system. Before you start, make sure you have:

• → A Artemis account (sign up free)
• → A robot or simulation running ROS 2 Humble or newer
• → Python 3.10+ on your robot's host machine
• → Network access from your robot to api.kairo.dev

Install the Artemis Agent

Install the Artemis agent on your robot host with pip:

pip install kairo-agent

Then authenticate using your API key from the dashboard:

kairo auth login --key YOUR_API_KEY

Next Steps