November 6, 2023

Release notes



  • The new List view in Catalog allows you to visualize data rows alongside data row fields such as data row ID, global key, external ID, row data, time of creation, annotations, predictions, and media attributes. You can also customize the list view by hiding columns and pinning columns.
  • The new performance dashboard uses a new timer system that solves the inaccuracy issues. It contains new filters and new measures such as rework, review times, and approval percentage to provide more insights into the performance of labelers/reviewers.
  • The improved Boost Express experience enables you to set up Boost Express in one seamless in-app experience. You can also set up project-specific labeling instructions in addition to the shared ontology instructions. To learn more, visit these docs.
  • The data ingestion pipeline has been revamped to enable fast and reliable data uploads capable of handling massive data volumes. To learn more about these improvements, read this blog post.
  • With the new Databricks pipeline creator, you specify a Databricks table to upload your data to Labelbox as a dataset. This is a no-code solution that is compatible with all Databricks tables assets stored on GCP, AWS, or Azure, as well as raw text stored directly in Databricks. To learn more, visit these docs.
  • We now offer Multi-factor authentication (MFA). If MFA is enabled, you'll be prompted to enter a code to complete the sign-on process. This code is generated by an authenticator app associated with your credentials. To learn more about MFA, read these docs.


  • In Catalog, thumbnails for some images were being rendered as text assets. Thumbnails for images are now reliably rendered as image assets.
  • When pasting a data row ID into the "Search your data" placeholder in Catalog, the results returned duplicate data rows. This issue has been fixed, and duplicate data rows are no longer returned.
  • Delete and re-queueing some labeled data rows was not working for newly created projects. This has been fixed.
  • For some text data rows, the detailed view showed the row data link rather than the text content. Now, the detailed view for all text data rows will show the text content.


  • By the end of 2024, we will block the reuse of ontologies containing the dropdown schema to new projects. To learn more about this deprecation, see this page.

Python SDK

The latest version of our Python SDK is v3.55.0. See our full changelog in Github for more details on what was added recently.

Version 3.55.0 (2023-11-06)


  • Fix the instantiation of failed_data_row_ids in Batch. This fix will address the issue with the create_batch method for more than 1,000 data rows.
  • Improve Python type hints for the data_rows() method in the Dataset.
  • Fix the DataRowMetadataOntology method bulk_export() to properly export global key(s).
  • In the DataRowMetadataOntology method update_enum_option, provide a more descriptive error message when the enum option is not valid.

Version 3.54.1 (2023-10-17)


  • Revised the notebooks to update outdated examples when using client.create_project() to create a project

Version 3.54.0 (2023-10-10)


  • Add exports v1 deprecation warning
  • Create method in SDK to modify LPO priorities in bulk


  • Remove backoff library