Skip to content
Onshape FeatureScript MCP Server

Onshape FeatureScript MCP Server

Released
5 views

Summary

PTC's MCP server for Onshape: describe a custom CAD feature in plain language and have an AI write, insert, run and debug the FeatureScript until it works.

Features

  • Generates FeatureScript for custom parametric CAD features from plain language
  • Inserts code into an Onshape document, runs it and reads the result
  • Iterates on errors until the feature behaves as described
  • Finished features run natively with no recurring AI token cost
  • Streamable HTTP endpoint usable from Claude, ChatGPT and Gemini

Installation

Set up this MCP server in your favorite AI agent — copy a ready-made configuration below.

Any MCP-compatible agent

Most agents (Claude, Cursor, Windsurf, VS Code, and more) read a standard mcpServers configuration.

  1. Open your agent's MCP configuration file.
  2. Merge the snippet below into it.
  3. Restart the agent — the "Onshape FeatureScript MCP Server" tools will be available.
{
  "mcpServers": {
    "onshape-featurescript-mcp": {
      "command": "claude",
      "args": [
        "mcp",
        "add",
        "onshape-featurescript",
        "--transport",
        "http",
        "https://fs-mcp.labs.onshape.app/mcp"
      ]
    }
  }
}

Description

An unusual MCP server: this one puts a CAD kernel behind the protocol. Onshape is PTC's cloud-native CAD and PDM platform, and FeatureScript is its open language for authoring custom parametric features — the mechanism behind every built-in feature in Onshape, available to users who want one that does not exist yet.

Writing FeatureScript has historically been the preserve of a small group of engineers who learned the language. The MCP server changes the loop: an engineer describes the feature they want, and the AI client generates FeatureScript, inserts it into an Onshape document, executes it, reads the result or the error, and iterates until the feature behaves. Text becomes code becomes geometry, with the CAD kernel — not the model — deciding whether it worked.

The economics are worth noting. Once a custom feature exists it is a normal Onshape feature: it runs in the document without further AI token cost, and can be reused across parts and shared with a team. The AI expense is a one-off authoring cost, not a per-use one.

Connecting

The server is a streamable HTTP endpoint at https://fs-mcp.labs.onshape.app/mcp. You need an Onshape account signed in, and a subscription to the FeatureScript MCP from the Onshape App Store. It works with coding-capable LLM clients including Claude, ChatGPT and Gemini.

Availability

Delivered through Onshape Labs, PTC's early-access programme for experimental AI and automation workflows, launched July 2026 alongside AI agents, drawing validation, AI-assisted rendering and robotics simulation. Treat it as early access rather than a GA feature.

Related MCP Servers

New

Arm's official MCP server for porting and tuning software on Arm — semantic docs search, codebase migration scanning, container architecture checks and LLVM-MCA assembly analysis.

New

Official AntV MCP server that turns a description of your data into a rendered chart — 26 visualization types from bar and line to sankey, mind-map, fishbone and geographic maps.

MCP: Headroom

by Headroom Labs

New

Context-compression MCP server that shrinks tool outputs, logs, files and RAG chunks before they reach the model, cutting 60-95% of tokens on JSON payloads.

MCP: Engram

by Gentleman Programming

New

Persistent memory for coding agents as a single Go binary — SQLite and FTS5 behind MCP, an HTTP API, a CLI and a TUI, with no Node, Python or Docker required.

Browse all MCP servers →