AlbumentationsX

AlbumentationsX is a Python library for image augmentation in computer vision and the actively developed library of the Albumentations ecosystem. It builds fast, reproducible augmentation pipelines that transform images together with their masks, bounding boxes, keypoints and 3D data, so training sets can be expanded while every label stays aligned with the altered image. It installs with pip. The library is aimed at machine learning engineers and researchers training vision models for tasks such as classification, segmentation, detection and pose estimation. Alongside the library, the team offers Explore, a browser workspace for testing transforms, comparing supported targets and preparing image workflows. The products page also lists data labeling, model training and deployment as future product lines that are only collecting expressions of interest, so they are not available to buy or use today. Documentation and a blog are published on the same site, and the project is affiliated with NumFOCUS. The library is available at no charge under the GNU AGPL version 3 only, which permits commercial use but brings source sharing obligations for covered distribution and for remote interaction with a modified version. A commercial license from Albumentations LLC removes those obligations for the agreed use in proprietary software. It is quote based: the price follows the licensed scope, covering legal entities, teams, projects or products, package versions, internal, hosted or customer delivered deployments, and term. Support, SLA, warranties and indemnity are separate options priced only if agreed. No figures are published. Teams building closed source products or hosted services should review the AGPL terms before adopting the library, since a commercial agreement is the route to using it without sharing source.