Skip to content
Tresca

Bonsai-mcp

MCP serverCommunity

A blender-mcp fork with IFC tools for Bonsai: spatial structure, quantities, georeferencing, plan PNGs and BC3 budgets. By JotaDeRodriguez.

66 starsUpdated 2 months agoGitHub data from 5 Nov 2026

At a glance

Type
MCP server
Works with
Blender / Bonsai (IFC)
How it connects
Drives a running copy of the tool
Access
Can modify models
Tool versions
Blender 4.0 or newer with the Bonsai add-on
Operating systems
Windows, macOS
Needs a Blender / Bonsai (IFC) licence
No
Transport
Local (stdio)
AI clients documented
Claude Desktop, Open WebUI (through the Docker REST/OpenAPI mode)
Licence
MIT
Language
Python

What it can do

  • Reads IFC project information, entities by type, property sets, relationships and the full spatial and element hierarchy of a model open in Bonsai.
  • Calculates element quantities and exports IFC data to JSON or CSV, filtered by type or level.
  • Reads or sets IFC georeferencing from an EPSG code or custom CRS with eastings and northings.
  • Places IFC elements at given coordinates, exports plan views as PNG and writes a BC3 (FIEBDC-3) cost budget in Spanish or English.

Limitations

  • The README gives uv setup steps for macOS and Windows only.
  • Needs Python 3.12 or newer for the server.
  • The Docker image exposes the tools as a REST/OpenAPI service, not as an MCP endpoint.
  • Keeps the original blender-mcp code execution tool, which can run any Python in Blender.

Install

{
    "mcpServers": {
        "Bonsai-mcp": {
            "command": "uv",
            "args": [
              "--directory",
              "\\your\\path\\to\\Bonsai_mcp",
              "run",
              "tools.py"
          ]
        }
    }
}

Source: github.com/JotaDeRodriguez/Bonsai_mcp#claude-for-desktop-integration

MCP servers and skills run code on your computer. Review the source before installing.

AI-generated models, drawings and results must be checked by a qualified engineer before use.