yaAnnotator

Help

Everything you need to go from a folder of images to an exported, training-ready dataset. The same content ships inside the app under Help (⌘?).

Topics

Annotating

The window, the editor canvas, drawing boxes and polygons, the inspector, and organizing images and splits.

Model-assisted labeling

The built-in detectors, Smart Select click-to-segment, pixel masks, custom Core ML models, and batch detection.

Import & export

COCO, YOLO, CreateML, CVAT, and Pascal VOC in and out, video frames, augmentation, presets, and what each format preserves.

Analyze Annotations

The data-quality report: class balance, geometry, validity checks, duplicates, and one-click fixes.

Keyboard shortcuts

Every shortcut in one place: files, images, tools, canvas, and views.

Accessibility

VoiceOver, keyboard navigation, color and contrast, reduced motion, and Voice Control.

Getting started from scratch

When yaAnnotator opens with nothing else to show, the Welcome window offers your recent libraries, a new library, or this help. To label a folder of images:

  1. New Library (⌘N): choose where to save the .yaannotatorlib file. It holds all your projects.
  2. New Project (⇧⌘N): give it a name.
  3. Import Images (⌘I): select your images. They're referenced in place, so keep the originals where they are.
  4. Add labels: click Labels (⇧⌘L) and add each class with a name and color. The first nine get ⌥⌘1 through ⌥⌘9.
  5. Open an image: select it and press Return, or double-click.
  6. Draw: press B for Box (drag a rectangle) or P for Polygon (click each vertex, then close it). Or press S and let Smart Select trace an object from a single click.
  7. Assign a label: click a label chip, or press ⌥⌘1 through 9. The active label is reused for your next annotation.
  8. Refine: press V for the Select tool; drag box handles, or drag vertices in the Selection inspector. Undo and redo anytime (⌘Z / ⇧⌘Z).
  9. Export: open the Export pane in the sidebar (⌘E), pick a format, and choose a destination. See Import & export.

Getting started from an existing dataset

  1. New Library (⌘N), as above.
  2. Import Dataset: use the + button in the sidebar (or File ▸ Import Dataset…) and select your dataset folder. yaAnnotator auto-detects common formats like COCO, YOLO, or CreateML and creates a project for it, with splits maintained from the import.
  3. Open images, refine or add annotations, and export, exactly as above.

Everything autosaves into the library as you work. Images are always referenced in place, never copied, so if you move or rename the original files you'll need to re-import them.