Key definitions

Below are the key definitions that you will see in the product, API, and the docs.




A project is a labeling environment in Labelbox. Think of it as a factory assembly line for producing labels. The initial state of the project can start with raw data, pre-existing ground truth, or pre-labeled data.


A model is an environment to evaluate a specific AI model performance


The Catalog is a warehouse of all data within an organization


An Asset is a single cloud-hosted file such as an image, a video, or a text file.

Data Row

A Data Row is the container that houses all of the following information for a single Asset:

  • URL to your cloud-hosted file
  • Metadata
  • Media attributes (e.g., data type, size, etc.)
  • Annotation information
  • Attachments (files that provide context for your labelers)


A Dataset is a set of Data Rows


Nearly everything in Labelbox is strongly typed. Schema is the master blueprint of your training data. It contains Ontologies, Features, and Metadata.


An ontology is a collection of Features and their relationships. It is essential for data labeling, model training, and evaluation.


A feature is the master definition of the entity you want the model to predict or humans to label. There are two kinds of features: objects and classifications. A feature can have multiple deeply nested sub-features.


Metadata is information about data. There are two types of metadata: reserved keys (user cannot change) and custom (user-defined).


An annotation is an instance of a Feature.


A label is a collection of all annotations on a data row


An editor is a labeling tool designed to create, review and edit annotations

What’s Next
Did this page help you?