Formas launches Cartesian, an AI‑driven 3D modeling tool for architecture and product design
AI‑driven 3D creation without CAD learning curve
Cartesian by Formas introduces an AI‑powered 3D modeling platform aimed at architects and product designers.
The system claims to interpret spoken commands, sketches or imported images and instantly generate precise geometry.
Users can “say it, sketch it, drag it” and receive a fully formed model without learning traditional CAD tools.
Every generated component retains its own identity, allowing designers to edit a single chair while the surrounding room remains unchanged.
Precision, editability and BIM‑ready outputs
The platform delivers exact faces, edges and solids that can be inspected, measured and altered directly.
Unlike many mesh‑based generators, Cartesian builds native NURBS solids and clean topology, eliminating polygon approximations.
Named elements and explicit relationships are embedded in the geometry, enabling a seamless path into BIM workflows.
A preview example shows a skylit restaurant where the booths along the wall stay exactly positioned after the user redefines the layout.
The workflow accepts a photo, rough plan, sketch, scan or existing model as the starting point.
Once created, the model can be opened in SketchUp, Rhino or any compatible CAD application for further refinement.
File sizes for exported 3DM and STL formats vary by category, ranging from 0.7 MB for simple product models to over 200 MB for large‑scale urban designs.
Furniture assets such as chairs and tables export at 8.4 MB (3DM) or 57.1 MB (STL), preserving precise connections and material definitions.
Interior scenes, including full rooms captured from scans or plans, export at up to 63.0 MB (3DM) while retaining editable placement of each object.
Architectural elements like walls, floors, stairs and structural openings generate files as small as 11.5 MB (3DM) for complex building models.
Industry‑wide file support and use cases
Residential and urban design modules can represent entire neighborhoods or streets, with NURBS geometry that stays editable across the whole site.
Landscape and infrastructure models combine paths, planting, seating and underground utilities in a single connected representation.
Advanced use cases include gridshells, lattices and doubly curved surfaces produced as exact geometry rather than faceted meshes.
Large‑scale projects such as stadiums or bridges are coordinated as one exact model, facilitating clash detection and coordination.
Manufacturing outputs provide unrolled panels, CNC toolpaths and tolerance data, ready for fabrication without additional conversion steps.
The system also supports 3D printing by adjusting wall thicknesses to match printer requirements, delivering printable STL files directly.
Game developers can export places and objects for virtual worlds, keeping the same precision available for real‑world construction.
Across all categories, Cartesian maintains native export options for DWG, 3DM, SKP and IFC, covering the major standards used in BIM and CAD pipelines.
By delivering real solids rather than rendered pictures, the platform enables measurement, cutting, printing and machining directly from the model.
The claim “what you keep is never redrawn” highlights that unchanged components retain their original geometry, reducing the risk of cumulative errors.
Integration with existing tools means firms can adopt the AI engine without purchasing expensive desktop CAD licences.
The preview waitlist invites early adopters to test the workflow and provide feedback before a broader release.
Why This Matters
This digest was compiled from:
Share this digest
People Also Ask
- AEF‑1 framework introduced for independent AI auditors, with Xai, OpenAI and Anthropic signing on
AEF‑1 sets a baseline for independent AI audits, while major labs pledge embedded third‑party evaluators to boost safety transparency.
- A Remark from Laurie Voss
Laurie Voss predicts that falling code costs will make user research and product design the main expense in software creation.
- Curated Reading List for Open‑Source AI and Open Models
Nathan Lambert’s curated list gathers essential essays, reports, and data to quickly bring readers up to speed on open‑source AI and open models.
Share your thoughts
Reactions, corrections, or insights — all welcome.
