Label Studio 1.12.0 🚀Automate & Evaluate Labeling Predictions Using LLMs & ML Models
Release notes

Introducing Label Studio 1.8.2

Label Studio Team


We're excited to announce the launch of our latest version of Label Studio, which features contextual scrolling - the ability to synchronize audio and paragraph segments so that the text automatically scrolls with the audio and video playback. We’ve also included some improvements to our machine learning backend so that you can more easily integrate a machine learning model for pre-labeling, active learning, and more! Read on for more details.

Install the latest release to get these new features, bug fixes and an important security update.

Contextual scrolling

Having both audio and text available for annotators to label leads to higher-quality annotation, however, manually scrolling through the transcript while the audio or video plays can lead to fatigue and slower labeling velocity. With contextual scrolling, the transcript is synced with the audio or video so that as the media plays, is fast-forwarded, reversed, or stopped, the text transcript will automatically scroll to the new listening point. This will improve annotation accuracy and provide a much better user experience for annotators and reviewers.

Specific functionality includes:

  • Bidirectional sync support between audio and paragraphs includes seek, play and pause activities
  • Interactivity with main audio player including playhead, timestamps and timelines
  • Visual indication of the audio segment currently playing and corresponding paragraph for quick orientation of playback progress
  • Ability to disable automatic scrolling for full control over playback
  • Segmented paragraph play/pause button synchronization with main audio player
  • Support for manual scrolling for long paragraphs
  • Video, audio, and segmented paragraphs synchronized support

See it in action here:

Stability improvements to ML backend

We’ve made some significant bug fixes to our machine learning backend that make it easier than ever to connect and use a machine learning model with Label Studio to automate labeling tasks and dynamically retrieve pre-annotations from your model.

With a model connected to Label Studio, you can:

  • Pre-annotate data
  • Use active learning to select the most relevant data for labeling
  • Enable interactive (AI-assisted) labeling
  • Fine-tune models based on recently annotated data

To learn more about how to implement our ML backend, check out the README file on GitHub.

⚠️ Important Security Update

We’ve fixed a moderate security risk by introducing a more secure approach to identity provider callbacks. You can now configure SECRET_KEY via an environment variable, ensuring that you have a unique string that cannot be easily compromised. Please update your instance of Label Studio to enable this security update.

We're eager for you to try out these new features and updates, and look forward to your feedback.

As always, we are here to support you in your data labeling journey. Stay tuned for more updates from Label Studio!


The Label Studio Team

Related Content