M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
B

Bonsai MCP

Community
MCP integration for IFC via Blender and Bonsai.
How to install ↓ Compare GitHub source repository ↗
Category
Dev Tools #243 of 438
Stars
★ 65 Popular
Transport
stdio (local process)
Runtime
Python 3.12+ · Docker · Prebuilt binary
Credentials
No credential needed
License
MIT
Last commit
Tools
17
51FMRS · D

Bonsai-mcp is an IFC-focused MCP server that leverages Blender to provide a rich set of IFC tools. Its notable features include georeferencing and BC3 budget export. It requires a Blender environment and is ideal for BIM professionals. Security consideration is advised with the execute_blender_code tool. Overall, it is a comprehensive solution for IFC integration.

Strongest · Documentation 13/20 Weakest · Reliability 7/20

Reliability
7/20
Security and permissions
10/20
Maintenance
10/20
Documentation
13/20
Setup experience
11/20
Why each score
Reliability 7/20
The code structure is clear, defining an MCP server and tools, but there are notable risks: the execute_blender_code tool allows arbitrary code execution, and the README describes features (like Docker REST API) that are inconsistent with a typical MCP implementation. There are no tests or CI configuration to verify tool behavior, so reliability cannot be scored above 12 due to lack of verifiable execution evidence. Score is 7.
Security and permissions 10/20
The most critical issue is the execute_blender_code tool, which allows remote execution of arbitrary Python code, enabled by default without confirmation or permission restrictions. This poses a severe security risk and could be abused for malicious actions. While it is an explicit feature, the lack of sandboxing or confirmation mechanisms lowers the security score. Other tools like export_ifc_data and export_drawing_png may write to the file system, but no clear issues are seen. According to red-line criteria, executing arbitrary code without confirmation is a serious security vulnerability, so security is scored 6.
Maintenance 10/20
The repository appears active with 58 stars and 1 open issue, but no commit or release information is provided. License is MIT, but there is no security response channel. Maintenance activity is inferred only from public info, lacking clear evidence, so score is 10.
Documentation 13/20
The README is detailed, covering installation, usage, troubleshooting, and configuration, but has some inconsistencies, such as the Docker section claiming REST/OpenAPI endpoints, which contradicts the MCP server functionality. Documentation lacks sufficient warnings about security risks, especially for execute_blender_code. Overall documentation quality is good but flawed, so score is 13.
Setup experience 11/20
Installation steps are clear, supporting multiple methods (MCP client, Docker, Blender add-on), but require pre-installed dependencies like Blender and Bonsai BIM, and configuration is somewhat complex. The Docker option adds an extra deployment path but also complexity. There are no automated tests or CI to verify installation, so per static calibration, setup cannot exceed 15; actual score is 11.

Static review · not runListed 2026-08-07

Read the FMRS scoring method →

Fit and risk

What it can accessWrites / deletes local filesRuns commands or code

Best for

  • Architects, engineers, and BIM professionals needing IFC model access.
  • LLM developers who want to interact with IFC data.
  • Automating design reviews and documentation generation.

Not for

  • Users not using Blender or the Bonsai BIM addon.
  • Processes requiring full automation without human interaction.
  • Users unfamiliar with IFC models (learning curve).

Required permissions

  • Execute commands on a local Blender instance via the addon server.
  • Execute arbitrary Python code in Blender (execute_blender_code).
  • Access and modify local IFC files.
  • Capture Blender viewport images and export data files.

Risks and side effects

  • The execute_blender_code tool can run arbitrary code, potentially causing data loss or unwanted changes.
  • Performance issues with large IFC models.
  • Misconfiguration in Docker deployments, leading to connection failures or CORS issues.

Setup

Before you start

Runtime：Python 3.12+ · Docker · Prebuilt binary

Other optional settings (4)
BLENDER_HOST optional Host of the Blender addon socket server when running via Docker, typically host.docker.internal; only needed for the Docker run mode.
BLENDER_PORT optional Port of the Blender addon socket server, e.g. 9876 in the example; only needed for the Docker run mode.
MCP_HOST optional Host the MCP server binds to, e.g. 0.0.0.0 in Docker; only needed for the Docker run mode.
MCP_PORT optional Port the MCP server listens on, e.g. 8000 in the Docker example; only needed for the Docker run mode.
  1. Install prerequisites: Blender 4.0+, Python 3.12+, uv package manager, and Bonsai BIM addon.
  2. Clone the repo: git clone https://github.com/JotaDeRodriguez/Bonsai_mcp.
  3. Install addon.py as a Blender addon.
  4. Configure the MCP client (e.g., Claude Desktop) with uv --directory <path> run tools.py. Alternatively, use Docker: build the image and run the container, exposing port 8000.

Check that it works

After enabling the Blender addon and clicking "Connect to Claude", a hammer icon and the IFC tools such as get_ifc_project_info should appear in the client. Asking "What is the basic information about this IFC project?" and receiving project details confirms the connection works.

Troubleshooting

  1. Ensure the Blender addon server is running and the MCP server is configured.
  2. Verify that the Bonsai BIM addon is installed and an IFC file is loaded.
  3. For Docker, ensure Blender is running and the addon server is started.
  4. Large IFC models may time out; try simplifying requests or breaking them into smaller steps.
  5. CORS issues: API has CORS enabled by default for all origins, but clients may need to adjust settings.

Things to try

Once connected, you can ask your AI assistant things like:

  • List all the walls in this IFC model
  • Show me the spatial structure of this building
  • Export all wall data to a CSV file
  • Georeference this IFC model using EPSG:4326 at latitude 40.7589, longitude -73.9851

Tools 17

get_ifc_project_info read-only
Retrieves basic information about the IFC project, including name, description, and counts of entity types.
list_ifc_entities read-only
Lists IFC entities of a specific type with options to limit results and filter by selection.
get_ifc_properties read-only
Retrieves all properties of a specific IFC entity by its GlobalId or from selected objects.
get_ifc_spatial_structure read-only
Gets the spatial hierarchy of the IFC model (site, building, storeys, spaces).
get_ifc_relationships read-only
Retrieves all relationships for a specific IFC entity.
get_selected_ifc_entities read-only
Gets information about IFC entities corresponding to objects currently selected in the Blender UI.
get_user_view read-only
Captures the current Blender viewport as an image, allowing visualization of the model.
export_ifc_data writes
Exports IFC data to a structured JSON or CSV file, with options to filter by entity type or building level.
Show 9 more tools
place_ifc_object writes
Creates and positions an IFC element in the model at specified coordinates with optional rotation.
get_ifc_quantities read-only
Calculates and gets quantities (m2, m3, etc.) for IFC elements, with options to filter by entity type or selection.
get_ifc_total_structure read-only
Retrieves the complete hierarchical structure of the IFC model including spatial elements and all building elements.
export_drawing_png writes
Exports 2D and 3D drawings as high-resolution PNG images with customizable resolution.
get_ifc_georeferencing_info read-only
Retrieves comprehensive georeferencing information from IFC files including CRS, map conversions, true north, and site coordinates.
georeference_ifc_model writes
Creates or updates georeferencing information in IFC models, allowing EPSG codes, custom CRS, and map conversions.
export_bc3_budget writes
Exports a BC3 budget file (FIEBDC-3/2016 format) based on the IFC model, supporting multi-language output.
execute_blender_code writes
Executes arbitrary Python code in Blender (legacy feature from BlenderMCP).
sequential_thinking read-only
Step-by-step thinking tool for structured problem solving and complex analysis.

Use cases

Query and analyze IFC models using natural language.
Generate technical reports with spatial hierarchy, quantities, and building elements.
Export BC3 budget files for construction cost estimation.
Use sequential thinking for maintenance planning or energy efficiency analysis.

Supported clients

Claude Desktop
Open WebUIPartial support

Listed from the project's documentation, not tested by this site.

Overview

Bonsai-mcp is a fork of BlenderMCP that extends the original functionality with dedicated support for IFC (Industry Foundation Classes) models through the Bonsai BIM addon for Blender. It lets LLMs read and modify IFC files by connecting to a Blender addon server. The server provides 11 IFC-specific tools, plus the execute code tool from BlenderMCP and the sequential thinking tool. It consists of two components: a Blender addon (addon.py) that creates a socket server in Blender, and an MCP server (tools.py) that implements the protocol and connects to the addon.

Similar servers

Context7 80 · B

Upstash's official server providing up-to-date third-party library docs for AI coding assistants

★ 62.9k · Tools 2 Compare with this →
Compare with other MCP servers

Source revision 1d8478f5d235 Data synced 2026-10-12 Read the FMRS scoring method