kernelCAD - CAD runtime for agents
MCP serverAI & modelskernelcad lets your AI create and edit 3D CAD models by working with .kcad.ts source files. Designs run on the OpenCASCADE geometry kernel, and changes can be reviewed deterministically, so the same edits always produce the same result. Once added, your AI can draft parts, adjust them on request, and review each change with you.
Use kernelCAD - CAD runtime for agents in Claude, ChatGPT or Ahel Desktop
Free. Sign in, add kernelCAD - CAD runtime for agents and connect your AI. About a minute.
Also: Claude Code · Cursor · Codex
Then ask your AI: use kernelCAD - CAD runtime for agents
No key or account is asked for. Not verified by us yet: Ahel has not run a call through it.
Details
Available today. Use it from your connected AI after setup.
No key or account is asked for. Not verified by us yet: Ahel has not run a call through it.
After adding it, ask your AI to model a simple part as a .kcad.ts file and review the result together. The project's code lives at github.com/w1ne/kernelcad-web.
What your AI can do with it
- Create 3D CAD models as editable .kcad.ts source files
- Edit existing designs by changing their source
- Review design changes deterministically, so identical edits always give identical results
- Build solid geometry on the OpenCASCADE kernel
Install kernelCAD - CAD runtime for agents
The server’s own address, for the clients that take one directly. Or connect Ahel once and every client you use reads it from one address, with the account kept on Ahel rather than in each client’s config.
Claude Code
claude mcp add --transport http --scope user kernelcad-cad-runtime-for-agents 'https://mcp.kernelcad.com/mcp'Run it once in your project, then open /mcp to approve any sign-in the server asks for.
Claude Desktop
https://mcp.kernelcad.com/mcpAdd a custom connector in Settings, paste this address, and approve the sign-in.
Cursor
cursor://anysphere.cursor-deeplink/mcp/install?name=kernelcad-cad-runtime-for-agents&config=eyJ1cmwiOiJodHRwczovL21jcC5rZXJuZWxjYWQuY29tL21jcCJ9Open the link and Cursor adds the server at that address.
ChatGPT
https://mcp.kernelcad.com/mcpIn Settings, enable Developer mode, create an MCP app, and paste this address. Your plan and workspace must allow custom apps.
Codex
codex mcp add kernelcad-cad-runtime-for-agents --url 'https://mcp.kernelcad.com/mcp'Run it once, then sign in with codex mcp login kernelcad-cad-runtime-for-agents if the server asks for an account.
From the project's README
As published by w1ne/kernelcad-web in README.md.
kernelCAD turns mechanical intent into editable .kcad.ts source,
deterministic review evidence, and exportable manufacturing artifacts.
The core loop is source-first: design brief -> editable .kcad.ts ->
evaluated model -> deterministic validation -> guided revision -> export package.
kernelCAD is not trying to be a browser clone of Fusion, Onshape, or SolidWorks. Replicad/OpenCASCADE are the kernel layer. kernelCAD is the agent-first workflow layer above the kernel: deterministic CAD source, feature history, diagnostics, validation, variants, and human review.
Philosophy
- Agent-First: The primary interface is
.kcad.ts, CLI, and MCP. The UI reviews and steers generated designs. - Editable Source: Geometry is deterministic source code with explicit feature history, params, assemblies, and diagnostics.
- Validation-Centered: Useful CAD output needs deterministic review for constraints, hardware fit, clearances, and exportability.
- Workflow Over Primitives: Product progress is measured by intent-level workflows, not by copying every traditional CAD button.
Agent workflow
For agents and contributors, .kcad.ts source is the design source of truth.
Rendered PNG/MP4 files, STEP/STL exports, score JSON, and capture-run metadata
are generated evidence or deliverables. Prompt briefs, source files, and
provenance metadata remain source. Change source first, regenerate explicit
targets, and avoid broad artifact refreshes unless you are intentionally
rebuilding a demo or release bundle.
Use deterministic checks before visual judgment:
kernelcad evaluate model.kcad.ts
# then run review_cad from an MCP client connected through `kernelcad mcp`
kernelcad export step model.kcad.ts -o model.step
kernelcad export stl model.kcad.ts -o model.stl
When a rendered artifact is produced, inspect the image/video itself and report what is visible: proportions, interfaces, floating geometry, occlusion, camera framing, and whether required features are legible. Passing tests are not a substitute for visual evidence.
When visual evidence matters, keep a deterministic inspection bundle:
kernelcad render inspect model.kcad.ts model.inspect
The v1 bundle writes a manifest and canonical RGB views. It can also emit
machine-readable mask, depth, and normals channels:
kernelcad render inspect model.kcad.ts model.inspect --channels rgb,mask,depth,normals
Use --focus <names> or --hide <names> when a specific feature id or
assembly part needs isolated review. Keep richer channels in the same manifest
packet; do not replace the canonical RGB views.
For real hardware, prefer catalog/vendor geometry via lib.fromSTEP(...)
instead of fake placeholder boxes or cylinders. This keeps mounts, clearances,
and mechanism checks tied to physical components.
Features
- Headless Core: Run CAD operations in Node.js or Web Workers without a DOM.
- Agent API: kernelCAD APIs for params, assemblies, NURBS, SDF bodies, sheet metal workflows, introspection, and feedback loops.
- Review Cockpit: Browser-based 3D preview for inspecting generated designs and validation results.
- Standard Exports: STEP/STL generation.
- Robust Kernel: Built on OpenCASCADE, with Replicad kept as a kernel implementation detail where it is still used.
Get Started
npm install -g kernelcad
Drop this in bracket.kcad.ts:
const w = param('Width', 60, { unit: 'mm' });
const h = param('Height', 40, { unit: 'mm' });
const t = param('Thickness', 5, { unit: 'mm' });
const base = box(w, h, t);
const hole = cylinder(t + 2, 4).translate(w / 2, h / 2, -1);
return base.subtract(hole).fillet(1);
Run it:
kernelcad evaluate bracket.kcad.ts
kernelcad export stl bracket.kcad.ts -o bracket.stl
That's it. For agents: kernelcad mcp runs an MCP server with dynamic model
introspection tools. See SKILL.md (bundled with the install) for the full API
surface and authoring guide.
Use with Claude Desktop
kernelCAD runs as an MCP server, so Claude Desktop can drive it locally. After
npm install -g kernelcad, add an entry to your Claude Desktop config file:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json - Linux:
~/.config/Claude/claude_desktop_config.json
{
"mcpServers": {
"kernelcad": {
"command": "npx",
"args": ["-y", "kernelcad", "mcp"]
}
}
}
Restart Claude Desktop. The kernelCAD MCP tools (review_cad, model
introspection, export helpers, etc.) appear in the tools list.
For prompt-driven authoring against the hosted gateway, sign in at
app.kernelcad.com/connect and use the
generated kernelcad mcp --cloud --token <…> snippet. The hosted surface adds
generate_kcad_from_prompt (turn a natural-language brief into a .kcad.ts
source) on top of the introspection and review tools, and serves the
kernelcad-authoring skill as an MCP resource so Claude Desktop loads it
automatically on connect. Free tier: 3 LLM-generated parts per month;
introspection and review tools are unlimited.
A one-shot
kernelcad install --claude-desktopcommand that edits this config for you is on the next slice — until then the snippet above is the interim path.
Contributing
Web app + dev workflow (clone the repo, npm install, npm run dev) for contributors who want to hack on the kernel or the visual debugger:
git clone https://github.com/w1ne/kernelCAD-web.git
cd kernelCAD-web
npm install
npm run dev # web visual debugger at localhost:5173
npm run build:cli # build the CLI bundle into dist/cli/
npm run test # full vitest suite
npm run qc # quick quality gate (lint + typecheck + tests)
Attribution and referral links
Exported files name kernelCAD in their metadata field (STEP FILE_NAME, GLB
asset.generator, the STL header, a DXF or XML comment), and public viewer
pages show a small "Made with kernelCAD" link. Some tools return vendor links
that can carry a referral tag; results with a tagged link include the
disclosure "Some links are referral links; kernelCAD may earn a commission."
A vendor with no tag gets the clean URL. Self-hosters and forks change
all of it in two files, or with KERNELCAD_REFERRAL_<VENDOR> and
KERNELCAD_REFERRALS=off. See ATTRIBUTION.md.
License
MIT — see LICENSE. The kernelCAD name is covered by the trademark policy.
Signals
- GitHub stars
- 25
- Forks
- 3
- Last commit
- Sep 2026
- Weekly downloads
- 202
Advanced
- Delivery
- kernelcad MCP server → your Ahel connector (mcp.ahel.ai) → your AI.
- Item type
- mcp-server
- Key
com-kernelcad-kernelcad- Source
- github.com/w1ne/kernelcad-web
- Hosted endpoint
https://mcp.kernelcad.com/mcp
More in AI & models
MCP server · upstash
More in AI & modelsAdButler
MCP server · adbutler
More in AI & modelsatom-mcp-server
MCP server · a7om-ai
More in AI & modelsottasia
MCP server · aiweather-anurag
More in AI & modelsunphurl
MCP server · 123ergo
More in AI & modelsmicrosoft-learn-mcp
MCP server · microsoftdocs
More in AI & models