zCAD AI: AI 3D that creates real parts, not just shapes

zCAD asks users to state a part's function, its dimensions, the interfaces it must meet, and any geometry that must stay fixed before generation. The company's showcase example is a roughly 14-part frame for a 5-inch FPV drone, built as a static assembly with no moving parts. Outputs carry driving dimensions, geometric constraints, and assembly mates, and the tool can check mating faces and clearances before a user continues the design. The category is getting crowded. In January 2026 Zoo shipped Zookeeper, a conversational CAD agent that can inspect, snapshot, and debug geometry as it generates production-oriented CAD. MecAgent also sells text-to-CAD for users of existing mechanical packages, and AdamCAD closed a $4.1 million seed round led by TQ Ventures. Autodesk has also announced its own generative model for editable 3D CAD, trained on Autodesk manufacturing datasets. For founders, editable parametric output is quickly becoming the baseline, so differentiation will depend on accuracy and workflow. One buyer comparison notes that production tolerances and GD&T still need a human. The free plan's terms also matter for anyone designing a data strategy: generation data used for training and free use is limited to non-commercial work, which affects how customer files and paid tiers should be structured.