Skip to content
Back to articles
open-source3 min

text-to-cad: the GitHub repo that lets AI agents design 3D parts ready to manufacture

KGA IT
2026-10-123 min
GitHubOpen SourceCADAgentsManufacturing
Capa: text-to-cad, plugin que dá a agentes de IA capacidade de CAD

What it is

text-to-cad (earthtojake/text-to-cad) is an open-source plugin that lets AI coding agents create real 3D parts. Instead of just writing code, the agent produces models as STEP, GLB, STL or 3MF, runs design-for-manufacturing (DFM) checks, generates engineering drawings and connects to 3D printing, sheet metal and CNC services.

Numbers checked via the GitHub API on October 12, 2026: ~19,200 stars, ~1,900 forks, MIT license, created April 2026 and updated daily. On October 11 it was among GitHub Trending's top repos with about 620 stars that day.

How it works

  • The CAD engine is the Python package cadgen, built on build123d and the Open CASCADE kernel, the same base as several open-source CAD tools.
  • It works with agents that support plugins or the skills standard: Claude Code, Codex, Cursor, Gemini and Grok, among others.
  • The plugin ships the skills plus a local viewer (MCP server cadgen mcp). In Claude Code models appear as cards in the conversation; in a terminal the agent returns a link to the browser viewer.
  • Recommended install: ask your agent to install it from the repo, or use the app's marketplace commands. Requires uv, Python 3.11+ and Node.js 20+.

A caution: telemetry

By default text-to-cad sends usage stats and crash reports, tagged with a random ID and, per the project, never including files, paths or prompts. To turn it off: uvx cadgen telemetry off. Companies with confidential designs should disable it first and review the policy.

Why it matters

CAD has always required expensive software and specialists. Tools like this let an agent describe a part in text, generate parametric geometry in code, validate it and send it to be made. For business readers:

  1. Fast prototyping: brackets, jigs, simple spare parts and enclosures can come out in hours.
  2. Human review still required: DFM checks help, but tolerances, materials and safety need an engineer.
  3. CAD as code: since the model is Python code, it can live in Git with reviews and versions, like software.
  4. Intellectual property: check where files go when using integrated fabrication services.

Sources


Want to use AI agents in engineering, prototyping or manufacturing? KGA IT offers a free consultation: kga-it.com.

Let's solve your technical challenges together.

KGA IT Solutions delivers AI, cloud, and DevOps expertise to address your specific challenges.

Contact Us