Onshape FeatureScript MCP Highlights How CAD Providers Must Use AI to Play to Their Strengths

Onshape, a cloud-native Computer-Aided Design (CAD) software company acquired by PTC in 2019, released FeatureScript MCP Server in August 2026. The capability is available through Onshape Labs, an early access program through which Onshape customers can explore new capabilities before wider releases.

FeatureScript MCP enables customers to connect Large Language Models (LLMs) directly to FeatureScript code to develop CAD automations, create CAD features, and tailor Onshape’s environment based on organizational or user-specific workflows. Features created using FeatureScript MCP are shareable across users, and version controlled with Onshape’s embedded Product Data Management (PDM) software.

FeatureScript MCP is a logical development to align the strengths of Artificial Intelligence (AI) with the advantages of Onshape. Onshape has two central strategic advantages relative to competitors: the platform is cloud-native, and FeatureScript is the foundational language used to build the platform’s tools. FeatureScript MCP enables users to leverage the well-understood strength of AI in writing code to create native User Interface (UI) tools that generate parametric geometry, rather than iteratively prompt AI to generate geometry that is less transparent and therefore harder to trust. Onshape’s cloud-nativity enables seamless sharing and control of features among different users, allowing customers to easily build and manage custom features.

The capability is not without its shortcomings: unless a company has an established enterprise AI, capitalizing on FeatureScript MCP requires using public LLMs such as Claude Code or Codex. Data leakage to public AI models will be a concern, as the top cyber risks among manufacturers are data breaches and new tech vulnerabilities such as AI, according to ABI Research’s Industrial and Manufacturing Survey 2H 2024/1Q 2025: State of Play for Digital Transformation (PT-3657). However, Onshape is undoubtedly playing to its strengths for leveraging AI in CAD, demonstrating a strategy other suppliers should follow.

For Onshape, the next steps must be continuing to leverage FeatureScript as a differentiator. To address data governance concerns among manufacturers, Onshape needs to develop its own FeatureScript coding agent that can be personalized for customers. Such a development would allow Onshape to maintain interoperability using its MCP, but offer a secure alternative for companies weary of exposing data to AI suppliers such as Claude and ChatGPT.

With developments to Onshape’s AI capabilities, PTC must also clarify Creo’s role in its AI roadmap, particularly for established manufacturing customers. Therefore, PTC should position Onshape as the trusted option for fast-moving, cloud-native AI experimentation, while Creo remains the primary choice for more controlled enterprise CAD deployments.

Competitors such as Autodesk and Dassault Systèmes are pursuing direct text-to-CAD capabilities with features including Autodesk Assistant and LEO from Dassault Systèmes. Siemens, however, remains the most conservative in using AI to generate editable parametric geometry.

Autodesk, Dassault Systèmes, and Siemens must approach AI-enabled generative design in two ways: automating the design of common, well-defined parts and exploring the design space to generate models under new constraints. The former is essential for delivering value to customers now, with outputs that are explainable and transparently generated. The latter is where companies can differentiate by combining native simulation engines with reduced-order models to help users interactively explore less-defined geometries as design requirements, constraints, and performance objectives continuously evolve.