CADSmith: Multi-Agent CAD Generation with Programmatic Geometric Validation

TL;DR

CADSmith is a multi-agent system that generates and iteratively refines CAD models from natural language using programmatic geometric validation and visual assessment, significantly improving accuracy and reliability.

Contribution

It introduces a novel multi-agent pipeline with nested correction loops that combine exact geometric validation and visual assessment for text-to-CAD generation.

Findings

Achieves 100% execution rate on a custom benchmark, up from 95%.

Improves median F1 score from 0.9707 to 0.9846.

Reduces mean Chamfer Distance from 28.37 to 0.74.

Abstract

Existing methods for text-to-CAD generation either operate in a single pass with no geometric verification or rely on lossy visual feedback that cannot resolve dimensional errors. We present CADSmith, a multi-agent pipeline that generates CadQuery code from natural language. It then undergoes an iterative refinement process through two nested correction loops: an inner loop that resolves execution errors and an outer loop grounded in programmatic geometric validation. The outer loop combines exact measurements from the OpenCASCADE kernel (bounding box dimensions, volume, solid validity) with holistic visual assessment from an independent vision-language model Judge. This provides both the numerical precision and the high-level shape awareness needed to converge on the correct geometry. The system uses retrieval-augmented generation over API documentation rather than fine-tuning, maintaining a current database as the underlying CAD library evolves. We evaluate on a custom benchmark of 100 prompts in three difficulty tiers (T1 through T3) with three ablation configurations. Against a zero-shot baseline, CADSmith achieves a 100% execution rate (up from 95%), improves the median F1 score from 0.9707 to 0.9846, the median IoU from 0.8085 to 0.9629, and reduces the mean Chamfer Distance from 28.37 to 0.74, demonstrating that closed-loop refinement with programmatic geometric feedback substantially improves the quality and reliability of LLM-generated CAD models.