OpenSCAD-MCP-Server

An MCP server (Python) that turns text prompts (and optionally images) into multi-view generated images, lets users approve/deny those images, reconstructs a 3D model using CUDA Multi-View Stereo (locally or via a remote GPU server on a LAN), and generates/refines/export parametric OpenSCAD-based outputs. Includes optional printer discovery/printing and a local web UI for the workflow.

Evaluated Mar 30, 2026 (22d ago)
Repo ↗ Ai Ml mcp openscad 3d reconstruction cuda remote-processing python zeroconf printer-discovery
⚙ Agent Friendliness
46
/ 100
Can an agent use this?
🔒 Security
40
/ 100
Is it safe for agents?
⚡ Reliability
32
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
55
Documentation
60
Error Messages
0
Auth Simplicity
55
Rate Limits
10

🔒 Security

TLS Enforcement
20
Auth Strength
40
Scope Granularity
20
Dep. Hygiene
50
Secret Handling
70

The README indicates use of .env for API keys and a shared secret key for remote CUDA MVS operations, which is a positive practice. However, it does not document transport security (TLS), auth enforcement on the main/local MCP server, key rotation, or detailed authorization/scopes. Remote LAN discovery increases the importance of network segmentation and access controls. Dependency hygiene and CVE posture cannot be assessed from the provided content.

⚡ Reliability

Uptime/SLA
0
Version Stability
40
Breaking Changes
30
Error Recovery
60
AF Security Reliability

Best When

Used on a trusted LAN (or isolated environment) by developers/technical users who can supply API keys and run OpenSCAD + (optionally) a CUDA MVS server with CUDA-capable hardware.

Avoid When

Avoid exposing the MCP/remote services directly to the public internet; avoid use when you cannot control network access to printer discovery/remote processing endpoints.

Use Cases

  • Text-to-3D model generation with multi-view reconstruction
  • Generating OpenSCAD parametric designs from reconstructions
  • Human-in-the-loop image approval before running expensive reconstruction
  • Offloading CUDA MVS reconstruction to a LAN GPU server
  • Exporting 3D assets in multiple mesh/parametric formats (OBJ/STL/PLY/SCAD/CSG/AMF/3MF)

Not For

  • Production deployments requiring strict security/compliance guarantees without additional hardening
  • Highly reliable unattended batch generation without monitoring/operational controls
  • Environments where running OpenSCAD, CUDA MVS, or GPU workloads is not feasible
  • Public-facing internet services without additional network/auth controls

Interface

REST API
No
GraphQL
No
gRPC
No
MCP Server
Yes
SDK
No
Webhooks
No

Authentication

Methods: API keys for Gemini/Venice image generation (via .env) REMOTE_CUDA_MVS_API_KEY for remote CUDA MVS operations LAN service discovery via Zeroconf (no user authentication described for discovery itself)
OAuth: No Scopes: No

Authentication is described only as shared API keys for remote operations; MCP/local server auth requirements are not clearly documented in the README. No OAuth, no fine-grained scopes, and no explicit auth scheme (e.g., bearer header name) is provided in the available documentation.

Pricing

Free tier: No
Requires CC: No

The project itself is MIT licensed; external API usage may incur fees.

Agent Metadata

Pagination
none
Idempotent
False
Retry Guidance
Documented

Known Gotchas

  • Remote processing appears to rely on LAN discovery (Zeroconf) and a shared API key; agents should ensure the correct server is targeted and avoid triggering unintended LAN requests.
  • CUDA MVS and OpenSCAD steps are heavyweight; failures may be due to missing CUDA builds, GPU availability, or system dependencies rather than MCP input validation.
  • The README does not specify MCP tool argument schemas or response schemas beyond example payloads; strict structured parsing may require additional verification.

Alternatives

Full Evaluation Report

Comprehensive deep-dive: security analysis, reliability audit, agent experience review, cost modeling, competitive positioning, and improvement roadmap for OpenSCAD-MCP-Server.

AI-powered analysis · PDF + markdown · Delivered within 30 minutes

$99

Package Brief

Quick verdict, integration guide, cost projections, gotchas with workarounds, and alternatives comparison.

Delivered within 10 minutes

$3

Score Monitoring

Get alerted when this package's AF, security, or reliability scores change significantly. Stay ahead of regressions.

Continuous monitoring

$3/mo

Scores are editorial opinions as of 2026-03-30.

8642
Packages Evaluated
17761
Need Evaluation
586
Need Re-evaluation
Community Powered