The monthly changelog showcases improvements in Roboflow over the past month. You can find the previous changelog here.

In February we had a major focus on reinforcing our foundation; after several months of rapid feature expansion we spent a considerable amount of time fixing bugs, improving infrastructure, and preparing for future initiatives.

Roboflow Organize

  • Added Upload Batches to keep track of which images were uploaded together.
  • Released the Unannotated Queue to improve labeling workflows.
  • Added an Upload Annotation API endpoint.
  • Refactored upload flow to save images before you start annotating so your work isn't lost if you leave the page or your browser crashes.
  • We now keep images in the same order they were uploaded in order to facilitate labeling video frames.
  • Improved team creation and invitations.
  • Added caching so counts don't need to be recalculated as often.

Roboflow Train

  • Scaffolding for on-device deployments (coming very soon).
  • Soft launched roboflow.js for browser-based realtime inference.

Roboflow Infer

  • Added support for static crop preprocessing option.
  • Updated to return prediction visualizations on the original image instead of the preprocessed image.
  • Improved error reporting.

Roboflow Annotate

  • Refactored Label Assist to re-use WebGL contexts for increased speed and reliability.
  • Reduced CPU usage and battery drain when annotating large datasets.
  • Added Label Assist onboarding flow.
  • Launched on Product Hunt! 🎉
  • Updated docs & marketing with info about Label Assist.