Getting Started

Sign Up and Create Your Account

To get started, sign up and create your first account by going to onyx-robotics.com and clicking “Sign In”. New users automatically get their own team with a Starter plan license and an example dataset to use and explore. Contact our team to upgrade your license plan.

Features Overview

The Onyx Engine consists of a web application and a user Python API. These products work hand-in-hand to enable controls engineers to train AI models on their hardware data.

The web platform is accessed via engine.onyx-robotics.com.

The Onyx Engine consists of the following main features:

  1. System overview and management

    1. Collaborate on your team’s system data via the node graph, table, and lineage notes.

  2. Dataset upload and processing

    1. Quickly upload data to view and gather insights.

    2. Process your dataset for AI model training.

  3. Model training and optimization

    1. Create and configure new AI models directly from your processed datasets.

    2. Run AI model training and optimization jobs.

    3. Analyze train loss and train predictions from your AI models.

Example Dataset

New teams can explore the platform via our provided example datasets. When logging into the platform, you will be greeted with the System Overview. The System Overview visualizes the full lineage of datasets and models in your system.

Overview

Explore the following actions:

  1. Click on an example dataset node in the graph to quickly view details via the Inspector panel on the right.

  2. Click on the name of the example dataset node in the graph or select the example dataset from the “Select” breadcrumb in the top bar to navigate to the dataset’s full object view to view and manage the data.

  3. Click on the “Table” tab to switch from the Overview to the Table. The Table contains a detailed list of all object versions in your system.

When you’re ready, upload your first dataset by clicking the “Upload” button in the bottom tool palette in the System Overview or the “Upload” button in the Table view. You can also drag in files directly in either view.