gNucleus AI targets engineering teams that want generative AI inside the CAD and simulation loop. Its core offering is a family of multimodal models trained on engineering data that can turn a text description, a sketch, a photo, or a legacy drawing (PDF/DXF) into editable, feature-based parametric CAD. Outputs land in native formats like CATIA, SolidWorks, and FreeCAD, plus neutral formats such as STEP and STL, so generated geometry fits existing workflows rather than replacing them.
Beyond single parts, the platform supports both top-down and bottom-up assembly generation with auto-created components and mating constraints. A suite of engineering agents extends the workflow: a Design Spec agent drafts parameters and constraints, a CAD agent produces parts and assemblies, a Simulation agent runs multi-physics workflows in tools like Ansys and Altair, and an Optimization agent interprets results and suggests improvements.
For enterprises, gNucleus offers proprietary model training and serving on your own engineering data, LoRA fine-tuning, managed or private cloud deployment, and professional data labeling and evaluation services for CAD and simulation datasets. It cites applications across automotive, robotics, and consumer electronics.
