freecad-tool-server

freecad-tool-server

Enables AI agents to programmatically build, analyze, and export 3D CAD geometry using FreeCAD through REST or MCP tools.

Category
访问服务器

README

FreeCAD Tool Server

An HTTP + MCP tool server that lets an AI agent of your choice drive FreeCAD programmatically — building, analyzing, and exporting real 3D CAD geometry through typed tool calls.

Agents call REST endpoints (or MCP tools) to create primitives, run boolean operations, loft airfoils, drill holes, fillet edges, check clearances, run finite-element analysis via Elmer, and export validated STEP files — receiving structured feedback at every step.

Your agent (Claude Code, Claude Desktop, any MCP/HTTP client)
      │  MCP tools  or  JSON over HTTP
      ▼
FreeCAD Tool Server  (main.py, typed models in models.py)
      │  in-process import
      ▼
FreeCAD 1.1 bundled Python  ──►  OpenCASCADE geometry kernel

The server runs inside FreeCAD's bundled Python interpreter — FreeCAD is imported as a module, not shelled out as a subprocess. That's why you launch it with FreeCAD's python.exe, not your system Python.


Prerequisites

  • FreeCAD 1.1+ installed (ships its own Python 3.11).
  • (Optional, for FEM) Elmer and Gmsh on your PATH for the /fem/* endpoints and elmer_*.py runners.

No system Python is required to run the server — FreeCAD's interpreter provides everything.


Setup

1. Install the Python dependencies into FreeCAD's interpreter

Windows

"C:\Program Files\FreeCAD 1.1\bin\python.exe" -m pip install -r requirements.txt

Linux / macOS (adjust to your FreeCAD install)

/path/to/FreeCAD/bin/python -m pip install -r requirements.txt

2. Start the server

Windows — edit the FreeCAD path in start.bat if needed, then:

start.bat

Linux / macOS — set FREECAD_PYTHON to your FreeCAD interpreter, then:

FREECAD_PYTHON=/path/to/FreeCAD/bin/python ./start.sh

The server listens on http://localhost:8000. Interactive API docs (Swagger UI) are at http://localhost:8000/docs.

3. Health check

curl http://localhost:8000/health
# {"status":"ok","freecad_available":true,"freecad_version":"1.1.1",...}

Connect your agent

Option A — MCP (Claude Code, Claude Desktop, any MCP client)

mcp_server.py exposes the tool server over the Model Context Protocol. Point your MCP client at it. For Claude Code, add it to your MCP config:

claude mcp add freecad -- "C:\Program Files\FreeCAD 1.1\bin\python.exe" /path/to/mcp_server.py

or add an entry to your client's MCP settings JSON:

{
  "mcpServers": {
    "freecad": {
      "command": "C:\\Program Files\\FreeCAD 1.1\\bin\\python.exe",
      "args": ["/absolute/path/to/mcp_server.py"]
    }
  }
}

The agent then gets every FreeCAD tool as a native MCP tool.

Option B — Claude Code reading AGENTS.md / CLAUDE.md directly

Clone this repo and open it in Claude Code (or any agent that reads AGENTS.md). The bundled AGENTS.md is a complete operating manual — the modeling loop, every endpoint, the common pitfalls, and the required integration checks. Start the server, tell your agent what to build, and it drives the HTTP API using that guide. A copy is provided as CLAUDE.md so Claude Code loads it automatically.

Option C — Raw HTTP from any language

Every tool is a plain REST endpoint. Minimal stdlib Python client (no extra deps) in tools/freecad_client.py:

from tools.freecad_client import FreeCADClient

fc = FreeCADClient()   # defaults to http://127.0.0.1:8000
fc.tool("POST", "/document/create", name="demo")
fc.tool("POST", "/shapes/add_box", name="b", length=50, width=30, height=10)
info = fc.tool("POST", "/model/get_shape_info", shape_name="b")
print(info["volume_mm3"])   # 15000.0

The core modeling loop

create_document
  → add primitives (add_box / add_cylinder / add_cone / add_wing)
  → boolean ops (boolean_union / boolean_cut / make_hole)
  → inspect (get_shape_info / bounding_box / check_interference)
  → finish (fillet_edges / chamfer_edges)
  → export (export_step)
  → validate (validate_step)   ← aim for is_clean == true, solid_count == 1

See AGENTS.md for the full endpoint reference, edge-selector rules, the in-place boolean-update pattern, and the mechanical integration checks (interference / clearance) that catch spatial errors before export.


Finite-element analysis (optional)

With Elmer and Gmsh installed, the /fem/* endpoints and the elmer_*.py runners cover a broad tutorial set — elasticity, heat, electrostatics, magnetostatics, fluid flow, eigenmodes, and more. These are heavier and require the external Elmer/Gmsh toolchain; the CAD tools work without them.


Repository layout

Path What it is
main.py FastAPI tool server — all HTTP endpoints
session.py FreeCAD session logic (geometry, booleans, export, FEM)
models.py Typed Pydantic request/response models
mcp_server.py MCP bridge — exposes the tools to MCP clients
freecad_bridge.py Headless FreeCAD discovery/import helper
AGENTS.md / CLAUDE.md Agent operating manual (identical content)
elmer_*.py Standalone Elmer FEM tutorial runners
tools/ Client helper, test suite, integration-check hook, example scripts
start.bat / start.sh Launchers (use FreeCAD's bundled Python)

License

MIT.

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选