August 18, 2021

Release notes



  • Annotation relationships (beta) in the Text editor are now available for all users with Text projects.

  • When you preview a Data Row from the Datasets tab or the Catalog tab, you have the option to view similar Data Rows in the carousel view.

  • Thumbnails for video assets in the Datasets tab will now appear as short-duration previews instead of still images.

  • You can now find these metrics on the project overview page:

    • Review metrics

    • Annotation count (replacing labelCountForWeek metric)

  • We added Python SDK helper snippets to the export page and the create dataset page in the app to offer suggestive use cases for command-line automation in Labelbox workflows.


  • Our text editor got some UI and performance updates.

  • You can manually resize the timeline in the video to make additional space in the canvas for the video being labeled.


  • Creating teams in Labelbox is no longer supported. The best way to divide up work amongst your labelers is to assign them to a project and specify their role (e.g., Labeler, Reviewer, etc). You can specify member roles at the organization level as well.

Python SDK

We released v3.0.0 of our Python SDK. Data processing dependencies are not installed by default for users that only want client functionality. To install all dependencies required for the new data modules (annotation types and MEA metric calculation) use pip install labelbox[data].


  • For details on which additional features are included in this latest release, see Labelbox SDK 3.0.0 Release.


  • In the latest version of the Python SDK (V3.0.0), the following have been replaced:

    • data_row.metadata() use data_row.attachments() instead

    • data_row.create_metadata() use data_row.create_attachments() instead

    • AssetMetadata (use AssetAttachment instead)


  • In the latest version of the Python SDK (V3.0.0), the following are no longer supported:


    • Project.create_prediction()

    • Project.create_prediction_model()

    • Project.create_label()

    • Project.predictions()

    • Project.active_prediction_model

    • DataRow.predictions

    • PredictionModel

    • Prediction