
California AI startup Illoca has launched the beta version of Plamo, a browser-based 3D modeling workspace designed to reduce the manual work involved in creating and modifying geometry. The platform converts natural-language prompts, sketches, drawings, and reference images into editable 3D models, tells AEC Magazine.
Illoca describes the approach as vibe modeling, drawing inspiration from vibe coding, where users communicate what they want while AI handles much of the underlying implementation. With Plamo, architects and engineers can focus on design intent while an AI agent translates their decisions into geometry.
A key feature is Plamo’s ability to generate editable boundary representation, or BREP, geometry supported by parametric logic. Unlike AI tools that produce static meshes or visualizations, Plamo allows users to change dimensions, relationships, and building systems as designs evolve. This makes the generated model an ongoing design environment rather than a fixed output.
Before modeling, Plamo’s agent interprets the user’s request using information from text, images, schematics, annotated drawings, or complete plan sets. Illoca says the system can reconstruct buildings, combine information from multiple references, add structural or MEP systems, and reorganize geometry according to project conventions. When information is missing or unclear, the agent can research details or request clarification instead of silently making assumptions.
Illoca sees applications extending beyond architecture. Engineers could use Plamo to develop structural and MEP systems within coordinated models, while developers and design-build teams could evaluate project alternatives earlier. Product and industrial designers could also transform sketches and requirements into parametric models suitable for downstream CAD and manufacturing workflows.
Plamo is currently available in beta with 1,000 free credits. Illoca has not yet announced additional pricing or supported CAD export formats. The startup, founded by former Autodesk AI Lab, Google DeepMind, and Tesla professionals, is backed by a $13 million seed round.
