AI 3D
NURBGen
Text-to-CAD: generate NURBS models from plain language.
★ 34 starsCustom
View guide →Give your AI agent CAD superpowers - generate real STEP/STL models from text
text-to-cad is an open-source ai 3d project by earthtojake (earthtojake/text-to-cad on GitHub). It has 18,426 GitHub stars and 1,830 forks, is released under the MIT license, and was last updated on GitHub on 8 October 2026.
Decide the part, dimensions and export format (STEP for CAD, STL for printing) you need.
Install it through your agent app - plugin directories for Codex/Claude Code, or a config entry for Claude Desktop.
The plugin starts the local CAD server via uv; make sure uv is installed.
Ask the agent for a part and review it in the CAD Viewer before exporting.
Your agent writes and runs CAD code (via the cadgen tooling) instead of producing a static mesh image.
Exports STEP/STP and STL files usable in downstream CAD and 3D-printing workflows.
Models appear as viewer cards you can orbit and inspect where the host app supports it, or via a browser link.
Documented installs for Claude Code, Claude Desktop, Codex, Cursor, Grok Build, Gemini and other MCP clients.
Ships skills that teach the agent the CAD workflow and conventions.
Generation runs locally through uv; your files stay on your machine (telemetry is optional and can be turned off).
Get the source code from GitHub and enter the project directory.
git clone https://github.com/earthtojake/text-to-cad.git
cd text-to-cadAdd the CAD server to Claude Desktop's config (Settings > Developer > Edit Config), then restart the app - the config pins a cadgen release run through uvx.
# Paste the server config from the README's Claude Desktop section, then restart Claude DesktopTelemetry can be disabled from the command line, and updates follow the per-host instructions in the README.
uvx cadgen telemetry offIt is maintained by earthtojake and contributors on GitHub.
Yes - text-to-cad is MIT licensed and free to use.
The CAD tooling runs locally for free; the agent driving it (Claude, Codex and so on) uses your own subscription or API access.
Compared with image-to-3D generators, text-to-cad produces code-based, parametric CAD with real STEP exports through your existing agent, rather than one-shot meshes.
Install it from your agent's plugin directory (or add the server config in Claude Desktop), restart the app, and ask for a simple part.
Text-to-CAD: generate NURBS models from plain language.
MAC (Multi-Agent CAD): A decoupled multi-agent framework for text-to-CAD generation via constrained test-time compute
Shap-E — OpenAI's text-to-3D and image-to-3D model: generate 3D objects as implicit functions and render them as NeRFs or textured meshes, MIT
Generate large-scale explorable 3D scenes with high-quality panorama videos from a single image or text prompt.
Text-to-3D that outputs editable parametric programs, not just meshes.
Modly — a desktop app that generates 3D models from images or prompts, running entirely on your own GPU.
The most-starred other AI 3D projects in the LibreAI catalogue — the realistic shortlist if text-to-cad is not the right fit.
Control Blender with natural language: an MCP plugin that lets any LLM model, light and animate in 3D
Pascal Editor: an open-source 3D architectural editor with MCP tools for humans and AI agents
Turn a reference image into a procedural, animation-ready Three.js model - as code
High-Resolution 3D Assets Generation with Large Scale Hunyuan3D Diffusion Models.
Official repo for paper "Structured 3D Latents for Scalable and Versatile 3D Generation" (CVPR'25 Spotlight).
Free and open — grab the source, star the repo, and start creating.