← All of Studio
Image annotator
Boxes, polygons, keypoints, and pixel-level masks.
Multi-class annotation with hierarchical taxonomies and instance tracking across a dataset. Built for computer-vision teams shipping detection, segmentation, and pose datasets.
Every geometry, one canvas
Boxes for detection, polygons for irregular objects, keypoints for pose, and brush-based masks for segmentation, switch tools without leaving the image.
- → Multi-class taxonomies with nested subcategories
- → Instance IDs that persist across an image sequence
- → Model-assisted pre-labels you accept, adjust, or reject
- → Zoom-to-pixel precision for small or dense objects
person · 0.97
car · 0.91
tree · polygon
Built for computer-vision datasets
Model-assisted pre-labeling
Point at a checkpoint and Studio pre-draws boxes or masks; annotators correct instead of drawing from scratch.
Instance tracking
Assign a stable ID to an object and carry it across a burst or sequence without re-labeling it each frame.
COCO / YOLO export
Export boxes, polygons, and masks in the formats your training pipeline already expects.