Product Update: June 2026

This month, Labellerr introduces major improvements to video event annotation workflows and foundation model-powered segmentation. These updates are designed to help annotation teams work faster, maintain higher accuracy, and manage complex video and image labeling tasks with greater efficiency.

Master Control Panel for Event Tagging Annotation

Managing event annotations across video frames can be challenging, especially when events span multiple frames and require frequent validation. To simplify this process, Labellerr now introduces Master Control for Event Tagging Annotation, a unified control panel that provides complete visibility and control over event annotations at the frame level.

With Master Control, annotators can:

• Update and modify event tagging annotations directly from a centralized interface.
• Correct event boundaries by adjusting the start and end frames of an event.
• Review which events are active or present in any selected video keyframe.
• Verify event annotations without navigating through multiple annotation panels.
• Improve annotation consistency across long video sequences.

This enhancement significantly reduces the effort required to manage temporal annotations and helps teams maintain higher-quality event labeling across large-scale video datasets.

Full SAM 3 Support with a Redesigned Annotation Experience

Labellerr now fully supports SAM 3 (Segment Anything Model 3) along with a completely redesigned UI/UX built for faster and more intuitive annotation workflows.

The new SAM 3 integration enables annotators to create high-quality segmentation masks with minimal manual effort while leveraging advanced visual and text-based prompting capabilities.

Visual Prompt Auto-Labeling

Annotators can now use existing annotations as visual prompts to quickly expand labels across similar objects in the same image.

• Select an already annotated object that represents the class or shape you want to propagate.
• Press "I" to initiate visual prompt labeling.
• SAM 3 analyzes the selected object and identifies visually similar instances in the image.
• All matching objects are automatically annotated with segmentation masks.

This workflow is especially useful when working with datasets that contain repeated items such as vehicles, tools, products, or other objects that appear multiple times in a single frame. Instead of manually tracing each instance one by one, annotators can use one labeled example to accelerate the rest of the task. This not only saves time but also helps maintain consistency in mask quality and object boundaries across similar instances.

One-Click Similar Object Annotation

With SAM 3 selected, Labellerr now supports a fast one-click workflow for annotating all similar objects in an image.

• Select SAM 3 from the model selection tab.
• Click on a target object in the image.
• Labellerr automatically detects and annotates all visually similar instances of that object.

This feature is ideal for scenes where the same object appears multiple times and manual instance-by-instance labeling would be time-consuming. By using the clicked object as a reference, SAM 3 can infer the remaining matching objects and generate segmentation masks in a single step. This helps teams handle dense or repetitive images more efficiently while reducing the chance of inconsistent annotations caused by manual tracing differences.

Text Prompt Segmentation

SAM 3 now supports text-driven object annotation directly within Labellerr, making it easier to label objects based on their semantic meaning rather than only visual selection.

• Select SAM 3.
• Choose the Text Prompt option.
• Enter the name of the object you want to annotate.
• Labellerr uses the prompt to identify the object and automatically segment it in the image.

Text prompting is especially valuable when annotators know exactly what they want to label but need a faster way to locate and segment it. This is useful for datasets with clearly defined categories, such as people, signs, containers, furniture, or other common objects. It also helps reduce the need for manual clicking or drawing when the object can be described directly in words. By combining semantic understanding with segmentation automation, text prompts make the annotation process more flexible and accessible for a wider range of use cases.

Why It Matters

These June 2026 updates focus on two key goals:

• Faster annotation workflows through SAM 3-powered automation and intelligent prompting.
• Improved video annotation management with centralized event tagging controls.

Together, these enhancements help annotation teams reduce manual effort, improve consistency, and accelerate dataset creation for computer vision and multimodal AI applications.