Automatically label image data
Label 1,000 images in under ten minutes. Auto Label uses state-of-the-art foundation models like Grounding DINO to help you label your image data. Over 2,000,000 images have been automatically labeled using Roboflow to date.
Reduce labeling time by 50% or more
In four steps, you can label thousands of images in hours, rather than days or weeks.
- 01
Set an ontology
Describe what you want to label in your images in natural language. Set as many labels as you want.
- 02
Evaluate your prompts
Visually inspect labels on a subset of your dataset. You can adjust confidence thresholds and class names to find a combination that works for your data.
- 03
Label your dataset
Once you have a configuration that works, you can label your data in the cloud.
- 04
Review annotations
Use Roboflow Annotate to review, remove, or adjust annotations from Auto Label, or work with Roboflow Labeling to outsource annotation review.
Review annotations with speed and accuracy
Use Roboflow Annotate to quickly review and modify annotations.
- Fast labeling workflows like keyboard shortcuts, intelligent defaults, and integrated label creation to annotate hundreds of images in minutes.
- Collaborate on annotation reviews with your team, allowing you to modify and approve annotations faster.
- Use model-assisted labeling powered by Segment Anything to add missing annotations with a single click.
