About Mcp Openscad Render Loop
MCP server that wraps OpenSCAD for parametric CAD iteration by agents and CAD users. It combines rendering, validation, parameter sweeps, and STL measurements in callable tools.
Key features
- 01Render STL with parameter overrides
- 02Create PNG previews
- 03Sweep parameter ranges
- 04Validate OpenSCAD syntax
- 05Extract module and function signatures
- 06Measure STL geometry
Use cases
- 01Render a CAD model with changed dimensions
- 02Compare outputs across parameter ranges
- 03Check source syntax before rendering
- 04Inspect dimensions and volume of an STL
MCP support
Exposes OpenSCAD rendering, validation, sweep, signature extraction, and STL metric tools to MCP clients.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEVerified asset on Archimedes Market. View the full 4-dimension Trust Report (security · quality · license · complexity) and the curated catalog on the asset page.
MCP OpenSCAD Render Loop
OpenSCAD wrapped as an MCP server so an agent can iterate on parametric CAD designs without bouncing in and out of the GUI. Render .scad source with parameter overrides, run parameter sweeps, validate syntax, extract module signatures, and compute STL geometry metrics.
Tools
render_stl— render a.scadfile or inline source to STL with-Dparameter overridesrender_png— preview render to PNG with camera + image-size controlparametric_sweep— Cartesian sweep over one or more parameter ranges, returns a manifest of outputsvalidate_scad— syntax check viaopenscad --check(no render)list_modules— extract top-levelmodule foo(...)andfunction bar(...)signatures with defaultscompute_metrics— bounding box, volume, surface area, triangle count from an STL
Why this server exists
OpenSCAD's CLI is well-designed but verbose. Most agent CAD workflows do the same loop: edit parameters → render → measure → adjust. This server collapses that to one tool call per iteration and returns geometry metrics directly so the agent doesn't have to parse STL bytes itself.
Quick start
# 1. Install OpenSCAD (must be on PATH)
brew install openscad # macOS
apt-get install openscad # Debian/Ubuntu
# 2. Install the MCP server
pip install mcp-openscad-render-loop
# 3. Set output directory and run
export MCP_OPENSCAD_OUTPUT_DIR="/tmp/scad-renders"
mcp-openscad-render-loop serve
Claude Desktop config:
{
"mcpServers": {
"openscad": {
"command": "mcp-openscad-render-loop",
"args": ["serve"],
"env": {
"MCP_OPENSCAD_OUTPUT_DIR": "/tmp/scad-renders",
"MCP_OPENSCAD_BINARY": "/usr/local/bin/openscad"
}
}
}
}
Typical agent workflow
Agent: "Design an enclosure for a 60x40x15mm PCB with 2mm walls and M3 mounting posts."
↓
1. list_modules on enclosure.scad → finds module enclosure(pcb_l, pcb_w, pcb_h, wall, post_dia)
2. render_stl with pcb_l=60, pcb_w=40, pcb_h=15, wall=2, post_dia=3 → enclosure_v1.stl
3. compute_metrics on enclosure_v1.stl → bbox 66x46x21, volume 18.4cm³, 1,240 triangles
4. parametric_sweep on wall ∈ [1.5, 2.0, 2.5] → 3 STLs + metrics → wall=2 is the strength/weight knee
5. render_png at iso camera → preview image attached to agent's reply
Safety
MCP_OPENSCAD_OUTPUT_DIRis the only writable path. Any output path outside this dir is rejected.- Each render is hard-capped at 60 seconds (configurable). Long sweeps run renders sequentially, not in parallel — keeps memory bounded.
- Inline
.scadsource is written to a temp file inside the output dir; nothing executes outsideopenscad. - No
import("/some/path.dxf")traversal — relative paths in.scadare resolved against the output dir or the source file's dir, never against/.
License
MIT.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.