OPEN LINK ↗
// 1h agoOPENSOURCE RELEASE
Labelme Makes Computer-Vision Labeling Local
Labelme is an open-source Python desktop tool for creating polygon, rectangle, circle, line, and point annotations, with exports for common computer-vision workflows. Its AI-assisted features use SAM, EfficientSAM, and YOLO-World to speed up labeling while keeping data on-device.
// ANALYSIS
Labelme remains a compelling, practical choice for developers who want hands-on dataset control without sending images to a labeling SaaS.
- –Supports segmentation, object detection, classification, and video annotation workflows
- –Local AI assistance turns rough clicks, boxes, or text prompts into editable annotations
- –Plain JSON files are easy to inspect, version, and convert into training pipelines
- –Its desktop-first workflow is simpler than full annotation platforms, but less suited to large distributed labeling teams
- –GPL-3.0 licensing and v7’s private internal modules matter for teams building custom integrations
// TAGS
labelmeopen-sourcevisiondata-toolsdevtoollocal-first
DISCOVERED
1h ago
2026-11-03
PUBLISHED
1h ago
2026-11-03
RELEVANCE
8/ 10
AUTHOR
GithubProjects