← Today's AI open source

Today's AI open source · 2026-11-13

tangency-labs/solidcraft

SolidCraft is a Rust-based parametric 3D CAD app with a built-in MCP server, so AI agents can do modeling too.

MCP

Repositorygithub.com/tangency-labs/solidcraft ↗
Main languageRust
LicenseApache-2.0
GitHub stars · forks★ 15,992 · 2,190 forks (as of 2026-11-13)
Created · last update2026-10-07 · 2026-11-12
GitHub topics3d-modeling, brep, cad, egui, mcp, parametric-cad, rust, step

Overview

SolidCraft is an open-source parametric 3D CAD app written in pure Rust. You apply constraints to sketches and build solids on a feature timeline. Changing a single dimension recalculates the features tied to that value. It runs both as a desktop app and in the browser.

With its built-in MCP server, compatible AI agents can run modeling commands, measure geometry, change parameters and export files. Designs created by an agent can be opened in the SolidCraft editor and edited further. Geometry computation is handled by SolidCore, a kernel ported to Rust from SMLib, NVIDIA's open-source C++ geometry library.

It suits developers who want to try having AI agents design mechanical parts or enclosures. Makers and engineers looking for free parametric CAD can also use it. The AI connection is optional, so it can be used as regular CAD without an agent.

Why it is worth a look

This is an example of MCP extending into CAD modeling work. What the agent produces is not a one-off mesh. It stays as a design file that keeps its parameters and feature history, so a person can continue editing it, which stands out.

Key features

What it is used for

Getting started

  1. Download the installer for Windows, the disk image for macOS or the AppImage for Linux, or install with sudo snap install solidcraft.
  2. To try it without installing, open the browser version on the website.
  3. To connect an AI agent, follow the MCP setup guide on the website to attach a compatible agent to the built-in MCP server.

What to weigh

Related terms and companies

Sources

Written with AI assistance from the repository README and GitHub data. Star counts and other numbers are GitHub values as of the date shown; features and licensing can change, so check the repository.