LAD

What it does

LAD is a SolidWorks add-in that generates sketches, features, assemblies, and macros from text descriptions. Users describe what they want to create in plain language, and the tool uses large language models to translate that instruction into SolidWorks geometry and commands. The add-in runs inside SolidWorks rather than as a separate application.

Who it is for

CAD users working in SolidWorks who want to speed up the initial design phase. This targets engineers and designers who spend time manually creating basic geometry or parametric features. The founders frame it as similar to how code-generation tools like Claude Code and Cursor have changed software engineering.

Pricing

The site does not show prices.

How it stands out

General-purpose LLM integration in CAD is still nascent. Most CAD vendors have not shipped native conversational design tools. LAD positions itself as filling a gap that SolidWorks, Fusion 360, and other major systems have not addressed. The use of LLMs for CAD workflows parallels the success of code-generation tools, suggesting there may be a real user need for reducing repetitive design tasks.

The fact that it works as a native add-in means users stay inside their existing SolidWorks environment rather than switching contexts. This could reduce friction compared to external AI design tools.

What a founder should check

First, verify the actual time savings in real workflows. The comparison to code-generation tools is intuitive, but CAD geometry requires precise constraints and tolerances. Check whether generated sketches typically need rework, or if the LLM reliably produces production-ready features on first try. Testing on varied design types (sheet metal, assemblies, complex surfaces) matters—a tool that excels at simple rectangular boxes may struggle elsewhere.

Second, understand SolidWorks' own AI roadmap. Autodesk owns both SolidWorks and Fusion 360, both of which could ship native LLM features. The switching cost for users is minimal if SolidWorks ships equivalent functionality. Look at whether Autodesk has filed patents or made acquisitions in generative design, and whether they are licensing LLM capability through existing partnerships.

Third, assess the moat around SolidWorks integration itself. Building an add-in is straightforward for vendors with API access. A sustained advantage likely depends on the quality of the LLM training data, the ability to learn from user corrections, or proprietary understanding of CAD workflows. Check whether LAD can meaningfully improve its model over time or if it will face head-to-head comparison with a native SolidWorks feature built by Autodesk's larger team.