Tech Soft 3D, the world leader in providing engineering software development toolkits (SDKs), announces the full release of HOOPS AI, the first framework purpose-built to unlock AI and machine learning for CAD data, following a successful beta program involving more than 30 companies.
HOOPS AI is now available to the general public, ready to support developers and data scientists looking to integrate 3D engineering data into modern machine learning pipelines.
Despite the transformative effect of machine learning technology on many industries, its ability to leverage CAD data has remained elusive. Existing tools rely on fragile workflows and improvised scripts, proving prohibitively expensive and time-consuming for most organizations. HOOPS AI provides the first framework to build machine learning models specifically with CAD data.
First introduced in late 2025 as a technology preview, HOOPS AI has built on the feedback from companies across the engineering software ecosystem. This helped shape the framework’s workflow, tooling, and machine learning integrations, ensuring the official release meets the needs of both software vendors and engineering teams working with complex CAD datasets.
Tech Soft 3D has processed more CAD data than any other CAD data access library provider, uniquely positioning them to develop and provide this technology. The company brings its decades of experience providing unrivaled CAD data access performance and accuracy to more than 750 partners, including Ansys, Siemens, Hexagon, the NVIDIA Omniverse, Unreal Engine, Unity 3D, and hundreds more.
“The official launch of HOOPS AI marks an important step in Tech Soft 3D leading the effort to bring AI to engineering data,” said Gavin Bridgeman, CTO, Tech Soft 3D. “For the first time, developers have a framework designed specifically to make CAD data usable in modern machine learning workflows.”
With this technology, Tech Soft 3D seeks to unleash the powerful combination of AI and CAD data for both its existing software vendor partners and the engineering industry as a whole.
HOOPS AI Empowers Machine Learning for 2026 and Beyond
HOOPS AI supports every aspect of the machine learning lifecycle, from CAD access, dataset preparation, optimized encoding, visualization, and scaling tools, and much more. Data scientists and developers can leverage this toolkit to build automated, intelligent experimentation systems that run hundreds or thousands of model variations, track results, and improve continuously. The result is a few engineers doing the work of a much larger team in weeks, rather than months.
In practice, HOOPS AI enables developers to apply machine learning to CAD data for tasks like part classification, manufacturing feature detection, and similarity search across large design libraries.
With its full launch, HOOPS AI will continue to develop and improve. The development team is looking to expand Python Access to more HOOPS Exchange-backed data, specifically targeting PMI. The long-term vision of the product aims to capture the expert engineering knowledge captured in CAD models and unlock it across teams. While current examples of HOOPS AI in action are trained on public datasets, the goal is to support workflows trained on a partner's own private data, tailoring the machine learning program more closely to their needs.
