ArtisanCAD: An Industrial-Level CAD Agent with Expert-Grounded Knowledge Distillation
06:00 · July 8, 2026 · arXiv cs.AI RSS
Computer-aided design (CAD) for industrial components requires long-horizon procedural modeling, robust feature dependencies, editable parametric geometry, and production-grade B-Rep execution. Existing text-to-CAD methods have made promising progress in generating CAD programs from natural-language descriptions, but they still struggle when user prompts are ambiguous, underspecified, or only describe high-level design intent. They also rarely exploit expert procedural knowledge naturally available in industrial workflows, such as CATIA operation recordings, macro logs, drawing notes, and engineering descriptions. We present \algname, a skill-guided industrial CAD agent with expert-grounded knowledge distillation. The core of \algname is CAD intermediate representation (CAD-IR), an executable procedural representation that encodes parameters, ordered operations, MCP tool bindings, dependencies, generated entities, and verification rules. CAD-IR plays two key roles: it first serves as the carrier for distilling expert CAD procedures into reusable parameterized skills; then it provides a procedural scaffold that turns vague or intermediate-level prompts into complete executable CAD operations. \algname retrieves expert-derived skills, instantiates and revises CAD-IR, executes the resulting procedure through a dedicated CATIA-MCP backend, and uses multi-view visual feedback for iterative refinement, and finally generates production-ready B-Rep models. On the Text2CAD benchmark, CAD-IR improves generation from intermediate prompts by reducing mean Chamfer Distance from $14.83$ to $9.88$, showing its ability to bridge ambiguous textual intent and executable CAD construction. On four complex automotive components, CAD-IR enables expert CATIA recordings to be distilled into reusable skills, allowing \algname to generate editable CATIA-native B-Rep models for new variant requests.
Summary
ArtisanCAD addresses the gap between current text-to-CAD systems and the demands of industrial design, where models must support long-horizon procedural construction, editable parametric features, and production-grade boundary representation output. Existing approaches often falter with ambiguous or high-level prompts and seldom incorporate the procedural knowledge embedded in expert workflows such as CATIA macro logs, operation histories, and engineering notes.
The system centers on CAD intermediate representation, or CAD-IR, an executable procedural format that captures ordered operations, parameter schemas, tool bindings for the CATIA-MCP backend, feature dependencies, generated entities, and verification rules. CAD-IR first acts as a medium for distilling recorded expert procedures into reusable, parameterized skills. It then functions as a scaffold that expands underspecified natural-language requests into complete, executable sequences while preserving expert-derived ordering and constraints.
During operation, ArtisanCAD retrieves relevant skills, instantiates and revises the CAD-IR through an agent loop, executes the procedure via the CATIA-MCP backend, and applies iterative refinement based on multi-view visual feedback. The result is an editable, CATIA-native B-Rep model suitable for downstream manufacturing and modification.
On the Text2CAD benchmark, CAD-IR alone reduces mean Chamfer Distance from 14.83 to 9.88 for intermediate-level prompts, demonstrating improved translation of vague intent into valid geometry. In four real automotive component cases, expert CATIA recordings are converted into reusable skills that enable the agent to produce variant models without rebuilding procedures from scratch.
Why it matters
This research is highly relevant for the Dutch AI market, particularly for its strong high-tech manufacturing and engineering sectors (e.g., ASML, VDL, Philips). Researchers and advanced practitioners can leverage these text-to-CAD advancements to automate and optimize complex industrial design workflows in the Netherlands.
