AIAny
AI Agent2026
Icon for item

text-to-cad

A library of reusable agent skills that generate, inspect, and hand off CAD, robot-description, and fabrication artifacts. Exports STEP/STL/3MF, writes URDF/SDF/SRDF, slices meshes to G-code, previews files in-browser, and includes off-the-shelf STEP part lookup and benchmarks.

Introduction

Why this matters

AI agents are increasingly asked to bridge design intent and manufacturable artifacts — not just output text. This project packages that bridge as composable skills: plain-language or image prompts can produce CAD geometry, robot description files, validated slice outputs, and browser previews so agents can complete fabrication or simulation handoffs without manual CAD steps.

What Sets It Apart
  • Focused on agent workflows, not a single GUI: skills are small, focused capabilities (CAD generation, CAD Viewer, step.parts, DXF, URDF/SRDF/SDF, G-code, SendCutSend, Bambu Labs integration) so agents can chain responsibilities rather than rely on a monolithic plugin. This means easier composition into multi-step automation like generate→validate→slice→send-to-printer.
  • Outputs engineering-grade exchange formats: native STEP export plus STL/3MF/GLB means artifacts are immediately usable in CAD pipelines and manufacturing; URDF/SRDF/SDF support targets robotics simulation and motion planning handoffs.
  • Pragmatic developer ergonomics: a Skills CLI and provider plugins (examples for Codex and Claude Code) let you install only the skills you need; benchmarks and demo assets offer reproducible prompts and validation checks to compare prompt-to-artifact quality.
  • Real-world validation: the repo includes benchmark prompts (e.g., flanges, enclosures, impellers, planetary stages) and heavy assets via Git LFS so you can inspect typical outputs rather than toy examples.
Who It's For and Tradeoffs

Great fit if you want to extend LLM/agent systems to produce engineering artifacts — robotics researchers who need URDF/SDF pipelines, makers who want agent-driven 3D prints, or teams automating CAD-to-manufacture handoffs. The project has thousands of stars and active demos, making it easy to prototype agent-driven design flows.

Look elsewhere if you need a WYSIWYG CAD authoring application for manual design work, a turnkey enterprise PLM integration, or a no-code visual interface — this repository assumes familiarity with agent frameworks, command-line installs, and basic CAD/robotics formats. It also relies on local toolchains (slicers, CLI printers) for final manufacturing steps and stores heavy demo assets in Git LFS.

Quick facts

  • Created 2026-04-22; widely used demos and CLI installers; includes integrations for Bambu Labs printing and SendCutSend preflight checks.
  • Key outputs: STEP/STL/3MF, URDF/SRDF/SDF, validated G-code, GLB previews, and off-the-shelf STEP part lookup.

If you plan to embed CAD generation into an agentic pipeline (generate → validate → slice → handoff), this repo supplies modular skills and benchmarks to speed that integration while leaving low-level tool choices to you.

Information

  • Websitegithub.com
  • Authorsearthtojake
  • Published date2026/04/22

Categories

More Items

Hugging Face
AI Model2026

Open-weights LLM fine-tuned for phone-based voice agents that prioritizes low latency and reliable tool/function calling. Based on NVIDIA Nemotron 3 Nano (30B total, 3.5B active), supports very long contexts (262,144 tokens) and recommends temperature=0 with thinking disabled for deployment.

GitHub
AI Agent2026

A curated collection of production-ready Agent Skills that turn tasks—presentation production, image generation, local KB retrieval, article assembly, and web-design—into agent-loadable skill folders. Uses a SKILL.md contract, supports multiple agent runtimes (Claude Code, Cursor, Codex), and offers modular install paths with pinned release zips.

GitHub
AI Agent2026

Provides a ruleset and skills that make AI coding agents prefer the simplest correct implementation: reuse existing code, prefer stdlib/native features, and only write minimal new code. Cuts generated LOC, tokens, cost and time while preserving validation and safety.