Smlibpy

Smlibpy is a Python library for 3D modeling on the SMLib NURBS kernel that lets scripts build parts.

Smlibpy is a Python library for 3D modeling that runs on the SMLib NURBS kernel. It allows developers and AI agents to build 3D parts directly through code, combining solids, cutting shapes, and exporting files. The library reads and writes STEP, USD, DXF, STL, and glTF formats while keeping exact geometry. It comes with an MCP server so an agent can build parts, and includes wheels for Linux, macOS, and Windows for Python 3.10 through 3.14. Smlibpy differs from ClassCAD by relying purely on scripts rather than persistent feature histories or constrained sketches.

Smlibpy at a glance

Best for: Software engineer
Runs on: Python, Web, macOS, Windows, Linux, API, open source
Pricing: MIT + Apache 2.0; no account required
Released: October 2026
Category: AI coding tool
Website: smlibpy.classcad.ch

Key features

Build solids from primitives and sketches
Combine parts using union, difference, and intersection
Apply fillets, chamfers, shells, and offsets
Extrude, revolve, loft, and sweep profiles
Read and write STEP, USD, DXF, STL, and glTF
Include an MCP server for AI agents
Calculate exact mass properties and volumes

Use it for

Generating 3D models with Python scripts
Allowing AI agents to build and test CAD parts
Running automated batch jobs and CI checks
Exporting exact geometry to STEP and other formats

Alternatives to Smlibpy

FastMCP — The standard framework for building Model Context Protocol servers and clients.
AirLLM — Runs massive 70B+ parameter models on small 4GB consumer GPUs.
Cad Studio — Generates parametric build123d models and ISO drawings through an integrated coding agent.
SOMA — Design 3D models instantly using natural language and AI coding agents.
MTPLX — It runs local models twice as fast via native MTP speculative decoding.
Wasi Python — Runs untrusted AI-generated code safely inside an isolated WebAssembly sandbox.
Lobole — Drives the GUI, Python scripts, and AI agents through a single unified command layer.
MLX LM — Runs LLMs and fine-tuning natively on Apple silicon hardware.
LMDeploy — Accelerates inference with high-performance engines and quantization tools.
SolidCraft — It runs on the desktop, in the browser, via CLI, and through an MCP server.
Hackerscad — Combines code-based modeling with direct mouse editing and an AI agent MCP server.
Mesh LLM — Run large local models across multiple machines through one OpenAI-compatible endpoint.

All Smlibpy alternatives