AnyLabeling is a desktop data labeling tool that combines manual annotation capabilities with AI-assisted auto-labeling powered by YOLO and Segment Anything models.
The tool addresses the time-consuming nature of dataset annotation by integrating automatic labeling alongside traditional manual annotation. It merges the functionality of LabelImg and Labelme with an improved interface and adds AI-powered auto-labeling. Users can annotate images with polygons, rectangles, circles, lines, and points, while YOLOv8 handles object detection and the Segment Anything family of models—including SAM, MobileSAM, SAM 2, SAM 2.1, and SAM 3—enable instance segmentation. SAM 3 uniquely supports open-vocabulary segmentation driven by text prompts. The tool also includes text detection, recognition, and key information extraction labeling. Hardware acceleration is available through CUDA, CoreML, DirectML, OpenVINO, and vendor NPU providers.
Teams working on computer vision datasets should choose this tool if they need to reduce annotation time without sacrificing quality. It suits projects requiring both object detection and segmentation labels, particularly those that can benefit from semi-automated workflows. The combination of multiple segmentation model variants allows users to balance accuracy against computational resources—MobileSAM for lightweight deployment, larger ViT variants for higher precision, or SAM 3 when semantic understanding through text prompts adds value. The tool supports multiple languages and provides pre-built executables for Windows, macOS, and Linux, eliminating setup friction.
Development activity shows consistent maintenance with recent releases addressing reliability concerns around model loading failures and data persistence. The release workflow validates compatibility across Python versions and tests all platform-specific builds on native runners before publication. Model weights download automatically on first use, and the tool persists annotation state reliably across save and load cycles.
