Markora Annotate
Image Labels & Training Data
Only for Mac
Free
Mac
Your Model. Your Data. Your Workflow.
Markora Annotate is an image annotation tool for building datasets on your Mac. Open an image folder, draw objects, refine outlines, and save annotations. A three-pane workspace keeps your image list, canvas, and object properties in view.
Flexible drawing and precise editing
Create rectangles, polygons, circles, points, lines, line strips, freehand closed outlines, and rotated boxes. Drag objects or vertices to refine them. Duplicate, undo, and redo to reduce repetitive work. Freehand outlines and rotated boxes are saved as polygons.
Review images at your own pace
Search by filename and filter unfinished images, completed images, or pending candidates. Edits save automatically. For images with one object, enable automatic advance after drawing. For multiple objects, mark the image complete after reviewing all annotations.
Local AI assistance
The built-in detector supports 80 common object classes, confidence and class filters, and detection for one image or an entire folder. Predictions remain candidates until you accept them. Import compatible Core ML detection models or use point and box selection to generate foreground outlines locally. Correct candidate geometry and labels, and prioritize images with low-confidence candidates.
Connect your training workflow
Read and write LabelMe JSON. Import YOLO detection boxes for one image or an entire dataset. Import and export COCO JSON and Pascal VOC XML detection data. YOLO export creates training and validation directories, data.yaml, classes.txt, and a source manifest. The Dataset Workspace offers folder-wide class counts, backed-up batch renaming, possible duplicate box checks, and clues for missing annotations. Copy annotations from the previous image when dimensions match.
Local processing, no account required
Images, annotations, and model inference stay on your Mac. No registration is required. The app contains no advertising, analytics SDKs, or image upload feature.
Formats and scope
Supports PNG, JPEG, TIFF, BMP, and HEIC. Detection exports use axis-aligned bounding boxes; polygons and circles become enclosing rectangles. COCO and VOC compatibility covers detection boxes, not segmentation masks or keypoints. Whole-dataset YOLO import requires an ordered class list and does not automatically parse data.yaml.
Custom models must be Core ML detectors with image input and NMS output. Arbitrary .pt and ONNX files cannot be imported directly. Foreground outlines use Apple Vision, not SAM. FITS, video, pixel-mask editing, YOLO OBB, and in-app model training are not supported. AI predictions may miss or misidentify objects. Quality checks provide review clues, not a guarantee that every omission will be found.
Available in English and Simplified Chinese. Switch languages using the language menu.
Requires macOS 14 or later.
Developer: Hu Dongsheng (胡冬生)
Support: hudongsheng356@gmail.com
Ratings & Reviews
- This app hasn’t received enough ratings or reviews to display an overview.
A New Way to Launch
A brand-new launch animation introduces a dynamic blue-to-cyan gradient, giving Markora Annotation a more distinctive visual identity from the moment it opens.
Built for Large Datasets
Thumbnail caching, next-image preloading, and background folder scanning make browsing and switching images significantly smoother—even in folders containing thousands of files.
Open Models. Faster Annotation. Smoother at Scale.
The developer, 海平 陈, indicated that the app’s privacy practices may include handling of data as described below. For more information, see the developer’s privacy policy .
Data Not Collected
The developer does not collect any data from this app.
Accessibility
The developer has not yet indicated which accessibility features this app supports. Learn More
Information
- Seller
- 海平 陈
- Size
- 33.5 MB
- Category
- Productivity
- Compatibility
Requires macOS 14.0 or later.
- Mac
Requires macOS 14.0 or later.
- Mac
- Languages
- English and Simplified Chinese
- Age Rating
4+
- 4+
- Copyright
- © 2026 Dongsheng Hu. All rights reserved.