Text To Cad
Open-source agent skills that turn text or image prompts into CAD models, G-code, and robot files.
If you already drive Codex or Claude Code, this is one of the cheapest ways to turn a prompt into a real STEP file, a validated G-code job, or a URDF — and nothing leaves your machine. The skill split matters: `cad` for geometry, `dxf` for cut layouts, `gcode` with real slicer CLIs, `urdf`/`srdf`/`sdf` for robotics, plus `dfm` and `dfam-check` that cite the rule behind each finding. It is not a CAD app and won't replace Fusion 360 or Onshape for constrained assemblies with persistent design intent. Pick it for automation, repetitive brackets, and print prep; skip it if you need a GUI timeline, a constraint solver, or cloud collaboration. Note the Codex plugin path needs Codex 0.142.0 or
Verified 7d ago · liveness 65/100 · cite: rightaichoice.com/tools/text-to-cad
- Developers using Codex or Claude Code who want prompts to produce real STEP and G-code files
- Robotics engineers generating URDF, SRDF, and SDF descriptions for MoveIt and simulators
- Makers turning text or image prompts into 3D-printable parts from the command line
- Fabricators preparing DXF cut layouts and validating them before SendCutSend upload
- Users who want a standalone GUI CAD app with a feature timeline and constraint solver
- Beginners who aren't comfortable with the command line or AI agent workflows
- Teams needing cloud-hosted, multi-user collaboration or shared model editing
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Text To CAD if you need a GUI feature timeline, a constraint solver, or cloud collaboration — it's an agent-mediated, locally installed skill library, not a standalone CAD application.
Nothing here is metered — the library is open source, but you still pay your AI agent provider's own token costs for every prompt you run through Codex, Claude Code, or Grok Build.
The library is published as open source and free to install. The real spend sits outside it: your existing Codex, Claude Code, or Grok Build subscription, plus the physical print or cut job when a file leaves your machine. Compared with a full parametric CAD seat — Fusion 360 or Onshape — that's a very low entry point for automation work, and you don't take on a per-seat licence just to batch out brackets or validate a DXF. The trade is capability, not cash: no timeline editing, no constraint
In short
Text To Cad — Open-source agent skills that turn text or image prompts into CAD models, G-code, and robot files. Best for Developers using Codex or Claude Code who want prompts to produce real STEP and G-code files, Robotics engineers generating URDF, SRDF, and SDF descriptions for MoveIt and simulators, Makers turning text or image prompts into 3D-printable parts from the command line. Free to use.
What people actually say about Text To Cad — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
45 mentions across 2 sources (Hacker News, Lemmy) · researched Jul 3, 2026.
Average across the 2 sources that answered — each source counts once, not each post.
- +Open-source and free with permissive license for commercial use.
- +Generates STEP files compatible with most CAD tools.
- +Includes a built-in browser viewer for quick previews.
- +Installs as a plugin for Claude Code and Codex easily.
- +Can generate robot description files (URDF, SDF) from text.
- −Describing complex geometry in words is frustratingly imprecise.
- −Not suitable for professional engineering workflows.
- −Lacks the iterative editing capabilities of traditional CAD.
- −Output quality varies heavily with the LLM used.
- −No visual modeling interface—purely text-driven.
- • No hidden costs—fully free and open-source.
Viability Score
How well maintained and how widely used is Text To Cad? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →Key Features
- Generate and edit CAD models from text or image prompts, with STEP as the primary output
- Export models to STL, 3MF, and GLB formats
- Create 2D DXF profiles, templates, gaskets, and cut layouts from Python sources or CAD geometry
- Produce dimensioned engineering drawings as PDF with views, hidden lines, dimensions, hole callouts, and a title block
- Slice supported meshes into printer-profiled FDM .gcode using real slicer CLIs
- Write URDF robot files with links, joints, limits, inertials, and meshes
- Add MoveIt planning groups, end effectors, poses, and collision rules via SRDF
- Build SDF simulator models and worlds with frames, physics, sensors, and lights
- Preview CAD and robot files in a local browser viewer
- Find off-the-shelf STEP parts: screws, bearings, motors, connectors
- Validate DXF and STEP files before SendCutSend upload
- Review sheet metal, CNC, and injection molding parts with measured evidence and cited DFM rules
- Measure mesh printability: wall thickness, overhangs, support volume, build orientation (DfAM check)
- Dry-run, upload, and cautiously start local Bambu Lab print jobs from validated .gcode
- Install as a provider-native plugin for Codex, Claude Code, or Grok Build, or via the Skills CLI
About Text To Cad
Text To CAD is a free, open-source library of agent skills that gives AI coding agents focused workflows for CAD, fabrication, and robot description files. Install it with `npx skills add earthtojake/text-to-cad` and your agent — Codex, Claude Code, or Grok Build — can create and edit CAD models from plain-language or image requests, with STEP as the main output plus STL, 3MF, and GLB export options. The current release, 0.7.6, shows 16,536 installs. It is a bundle of narrow skills rather than one blob. `cad` handles the modeling; `dxf` produces 2D cut layouts, templates, and gaskets; `gcode` slices supported meshes into printer-profiled FDM files using real slicer CLIs; and `urdf`, `srdf`, and `sdf` cover robot structure, MoveIt planning, and simulator worlds. Other skills include dimensioned engineering drawings as PDF, a `dfm` reviewer for sheet metal, CNC, and injection molding that cites the rule behind each finding, a `dfam-check` that measures wall thickness, overhangs, and support volume, and `step-parts` for off-the-shelf components like screws, bearings, motors, and connectors. Everything runs locally. A bundled `cad-viewer` previews CAD and robot files, `sendcutsend` validation checks DXF and STEP before upload, and `bambu-labs` dry-runs, uploads, and cautiously starts local print jobs. Provider-native plugin installs exist for Codex, Claude Code, and Grok Build, though the Skills CLI is the recommended path for regular use. This is agent-driven design, not a parametric CAD replacement. If you live in Fusion 360 or Onshape for constrained assemblies and full timeline editing, those still do that job. Text To CAD is the practical, cost-free bridge for developers who already work with AI coding tools and want text to become physical parts.
Behind the Verdict
Text To CAD's value is its narrowness. Instead of one vague CAD assistant, it ships thirteen-plus skills each mapped to a job: `cad` creates and edits models from text or images and exports STEP, STL, 3MF, or GLB; `dxf` generates 2D profiles, templates, gaskets, and cut layouts from Python sources or CAD geometry; `gcode` slices supported meshes into validated, printer-profiled FDM files using real slicer CLIs; `urdf` writes robot structure with links, joints, limits, inertials and meshes, `srdf` layers MoveIt planning groups, end effectors, poses and collision rules on top, and `sdf` builds simulator worlds with frames, physics, sensors and lights. For documentation and checking there is `engineering-drawing` (PDF views, hidden lines, dimensions, hole callouts, title block), `dfm` (sheet metal, CNC, injection molding, with measured evidence and the cited rule behind every finding), and `dfam-check` (wall thickness, overhangs, support volume, build orientation). `step-parts` pulls off-the-shelf screws, bearings, motors and connectors into your assembly, and `sendcutsend` validates DXF and STEP before you upload. Print handoff is covered by `bambu-labs`, which dry-runs, uploads, and cautiously starts local Bambu Lab jobs. Strengths: local execution, no subscription, STEP as a real engineering interchange format rather than a mesh-only toy, and skills that cite their own rules instead of hand-waving. The `cad-viewer` skill gives you local preview without a cloud account. Install paths are flexible — `npx skills add earthtojake/text-to-cad` for the CLI, or provider-native plugins for Codex, Claude Code and Grok Build. Weaknesses are structural, not bugs. It is not a standalone application: everything is agent-mediated, and the source material states plainly that workflows need command-line and agent familiarity. There is no feature timeline, no constraint solver, and no persistent design intent across edits, so constrained assemblies that must stay parametric belong in a real CAD package. Multi-user collaboration and shared model editing aren't part of the picture either — it runs on your machine. The Codex plugin requires Codex 0.142.0 or newer and older versions skip it silently, so check your agent version before you blame the skill. Where it fits: developers and robotics engineers already living in an agent loop who want text to become a physical part without a licence or a cloud round trip; makers who want to go from a downloaded STL to a printer-profiled G-code job and straight onto a Bambu Lab machine; fabricators who need a DXF profile validated against SendCutSend's rules before upload. Where it doesn't: designers who want to click a fillet in a timeline, beginners with no CLI comfort, and teams that need shared cloud workspaces.
Researching Text To Cad? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Text To Cad actually fits — and what changes day-one when you adopt it.
Installs the library with `npx skills add earthtojake/text-to-cad`, then asks the agent to model a custom mounting bracket and export STEP alongside an STL preview.
Outcome: A real STEP file and mesh land in the local project folder, with the `cad-viewer` skill showing the part in the browser before anything is committed.
Describes a joint and link list, has the agent write a URDF with limits, inertials, and meshes, then layers MoveIt planning groups and collision rules on with SRDF and builds an SDF world for the simulator.
Outcome: A consistent robot description set — structure, planning, and simulation — generated from one prompt chain instead of hand-written XML.
Slices a downloaded STL into printer-profiled FDM G-code with the `gcode` skill, runs `dfam-check` for wall thickness and overhangs, then dry-runs and uploads the validated file to a local Bambu Lab printer.
Outcome: A printability-checked G-code job that only reaches the printer after a dry-run, which catches obvious failures before filament is spent.
Use Cases
- Generate a CAD model of a custom bracket from a text description and export it as STEP
- Create a URDF robot description from a list of joints and links, then add MoveIt planning groups as SRDF
- Slice a downloaded STL into printer-profiled G-code for a Bambu Lab printer
- Find and insert a standard off-the-shelf STEP part such as an M5 screw or a bearing
- Export a 2D DXF profile for laser cutting from parametric Python code
- Validate DXF and STEP files against SendCutSend rules before uploading
- Check mesh printability — wall thickness, overhangs, support volume — before committing a print
- Produce a dimensioned PDF engineering drawing of a finished part
Models Under the Hood
as of 2026-09-25
Limitations
- This is an open-source skills library that runs locally and is installed via the Skills CLI (`npx skills add earthtojake/text-to-cad`) into an AI agent, or as a provider-native plugin for Codex, Claude Code, or Grok Build.
- It is not a standalone application; workflows are agent-mediated, so you need some command-line and agent familiarity.
- There is no feature timeline and no constraint solver, so parametric relationships and design intent do not persist across edits the way they do in a dedicated CAD package.
- It runs on your machine, so there is no cloud-hosted multi-user collaboration or shared model editing.
- The Codex plugin requires Codex 0.142.0 or newer — older versions skip it silently, so check your agent version before troubleshooting the skill.
as of 2026-10-02
Verification history
We have re-verified Text To Cad 6 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Text To Cad tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Free
$0
Ideal for
Developers, robotics engineers, makers, and fabricators who already run Codex, Claude Code, or Grok Build and want CAD, DXF, G-code, and robot files without a per-seat licence.
What this tier adds
Starting tier — the full skill library (cad, dxf, gcode, urdf, srdf, sdf, step-parts, dfm, dfam-check, and more) installs locally at no cost, with your agent subscription and physical print or cut jobs as the only outside spend.
Where the pricing makes sense
The company stage and team size where Text To Cad's pricing actually pencils out — and where peers do it cheaper.
The library is published as open source and free to install. The real spend sits outside it: your existing Codex, Claude Code, or Grok Build subscription, plus the physical print or cut job when a file leaves your machine. Compared with a full parametric CAD seat — Fusion 360 or Onshape — that's a very low entry point for automation work, and you don't take on a per-seat licence just to batch out brackets or validate a DXF. The trade is capability, not cash: no timeline editing, no constraint
Setup time & first value
How long it actually takes to get something useful out of Text To Cad — broken out by persona, not the marketing-page minute.
Developers comfortable in a terminal: roughly 5–10 minutes — run the install command, confirm your agent is on a supported version (Codex 0.142.0 or newer for the plugin path), and restart the agent so new skills appear. Robotics and fabrication users: another 15–30 minutes to learn which skill handles URDF, DXF, or G-code, since the value is spread across a bundle of narrow skills rather than
Switching to or from Text To Cad
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From writing CAD scripts by hand in build123d or similar Python CAD libraries: keep the same Python-based output but let the agent draft and iterate the geometry from a prompt.
- →From a mesh-only AI 3D generator: switch to `cad` for STEP output, which is an engineering interchange format rather than a printable-only mesh.
- →From doing DXF layout work manually in a 2D tool: use the `dxf` skill to generate profiles and cut layouts, then check them with `sendcutsend` before upload.
- ↗To Fusion 360 or Onshape: move to a parametric seat when you need a feature timeline, a constraint solver, and persistent design intent across edits.
- ↗To a mesh-only AI 3D generator: move over if you only ever need a visual or printed mesh and don't care about STEP interchange.
- ↗To a hand-written Python CAD script workflow: move over if you want full manual control of every line rather than agent-drafted geometry.
Integrations
Resources & Guides
Tutorials & Learning
3D Printing AI Text-to-CAD 3D Model
3D Luke
Text-to-CAD AI is here! Use text prompts to create and manipulate 3d models, for those of you who
Sabrina Ramonov 🍄
Text-to-CAD: Creating 3D Models With Simple Text Prompts | Revolutionary CAD Design Tool
Champion Edtech
YouTube returned 6 videos for “Text To Cad”, and we withheld 3: 3 did not mention Text To Cad. Showing the 3 we can prove are about Text To Cad.
Official links
Tools that pair well with Text To Cad
Common stack mates teams adopt alongside Text To Cad, with the specific reason each pairing earns its keep.
Zoo
Zoo Design Studio is an AI-native CAD workspace where Zookeeper, a conversational agent, turns plain-language prompts into editable B-rep parametric models.
Tripo AI
Tripo AI turns text prompts, images, or sketches into engine-ready 3D models — quad-mesh game assets, print-ready STLs, and textured characters in minutes.
Rodin
Rodin turns text prompts and reference images into textured 3D models with bounding-box, voxel, and point-cloud ControlNet plus per-region Partial Edit.
Featured Head-to-Head Comparisons
Text To Cad vs Locus Robotics
Choose Text To Cad for free, open-source AI CAD generation from natural language if you are an engineer comfortable with CLI tools; pick Locus Robotics for a proven, scalable warehouse automation system (RaaS) that boosts productivity 2-3x, ideal for high-volume eCommerce fulfillment centers.
Text To Cad vs Truleo
Choose Text To Cad if you're an engineer or maker who needs free, open-source AI-assisted CAD and robot file generation from natural language, and you're comfortable with CLI tools. Choose Truleo if you're in law enforcement and need a paid, CJIS-compliant intelligence platform that stitches together siloed databases (RMS, CAD, jail calls, BWC) to surface leads and cut report-writing time. These tools serve unrelated domains—your decision hinges on whether you're designing parts or investigating crimes.
Text To Cad vs Presto Voice
Text To Cad and Presto Voice serve completely different niches, so the right choice depends on your domain. If you're an engineer or maker automating CAD generation, Text To Cad's free, open-source AI agent skills are a no-brainer. If you're a QSR chain optimizing drive-thru revenue, Presto Voice's proven voice AI with up to 95% automation and up to 6% revenue lift is a powerful investment. Choose based on your role and problem—text-to-CAD or drive-thru automation.
Alternatives to Text To Cad
View all
Zoo
Zoo Design Studio is an AI-native CAD workspace where Zookeeper, a conversational agent, turns plain-language prompts into editable B-rep parametric models.
Frequently Asked Questions
Best-of guides
Used Text To Cad? Help shape our editorial sentiment research.