Labelbox documentation

About the project configuration

At the very minimum, every project in Labelbox is required to have the following:

  • A Dataset

  • An Editor

  • An Ontology

These 3 configurations are the foundation for your labeling workflow. Once you have a Dataset, Editor, and Ontology configured for your project, you can further customize your labeling process with QA tools, queue customization, issues & comments, etc.

The relationship diagram below illustrates the general framework of a typical Labelbox project configuration.


To learn how to configure a project in the Labelbox UI, see How to create a project.

To learn how to set up a project via the Python SDK, see this Google colab tutorial.