X-AnyLabeling

X-AnyLabeling is a lightweight, cross-platform desktop application for annotating images, video, text, point clouds, and multimodal data — supercharged with state-of-the-art AI models that do most of the labeling for you. It unifies manual annotation tools with one-click AI assistance from models like SAM, YOLO, and Grounding DINO in a single, fast interface.

Overview

Labeling data is the bottleneck of most computer vision projects, and X-AnyLabeling attacks it from both sides: precise manual tools for polygons, rotated boxes, keypoints, and OCR, plus AI auto-labeling that proposes annotations you simply review and correct. It supports classification, detection, segmentation, pose estimation, matting, and more, and exports to the formats training pipelines expect, including COCO, YOLO, and Pascal VOC. The application runs on Windows, macOS, and Linux, supports GPU acceleration through ONNX Runtime, and can also run in a server mode for team workflows. The project has more than 10,700 GitHub stars, is actively maintained with updates in October 2026, and is GPL-3.0 licensed.

Key Features

• AI-assisted labeling with SAM, YOLO, Grounding DINO, and more
• Annotation for images, video, text, point clouds, and multimodal data
• Supports detection, segmentation, pose, OCR, matting, and rotated boxes
• Exports to COCO, YOLO, VOC, and other standard training formats
• Cross-platform desktop app with GPU acceleration via ONNX Runtime
• Video tracking and interpolation to speed up sequence labeling

💳 Pricing

X-AnyLabeling is free and open source under the GPL-3.0 license, with no paid tiers or feature locks.

Who Is It For

Computer vision teams, data-labeling operations, and researchers who need fast, AI-accelerated annotation across many data types and tasks.