Vibeleaderboard
Index / tool
Visit github.com
Category
AI Agents
Rank
Pricing
Open Source
Type
TOOL
Use case
Design & Media
Interfaces
MCP · Host Plugin
Builder
ahujasid
Date

About

Connects Blender to Claude AI through the Model Context Protocol, allowing Claude to directly create, modify, and control 3D models, scenes, and materials in Blender through natural language prompts. Enables AI-assisted 3D modeling with two-way communication between Claude and Blender.

What it does

BlenderMCP pairs an in-application Blender add-on with a Python MCP server. The server translates an AI client's requests into JSON commands over a socket, while the add-on executes them inside Blender and returns scene data, results, or viewport images.

Why it's ranked here

The architecture is direct and unusually capable: it combines scene inspection, asset retrieval, viewport screenshots, material work, and arbitrary Python execution. That breadth makes it useful for serious experimentation, but installation friction, GUI dependence, telemetry terms, and unrestricted code execution demand informed use.

What's good

It can inspect before acting, which gives the model feedback about scenes and individual objects. A connection lock prevents overlapping commands from mixing responses. Poly Haven, Sketchfab, Hyper3D, and Hunyuan3D integrations extend the workflow beyond primitive generation, while stored credentials support repeated use.

Tradeoffs

Setup requires both a Blender add-on and a separately configured MCP server. GUI-launched clients may need absolute executable paths, and only one server instance should run. Blender background mode is explicitly unsupported. Commands can wait three minutes before timing out. Arbitrary Python execution expands capability and risk. Telemetry is enabled unless disabled, and the terms permit broad future uses of collected data when consent is given.

How to use it well

Use it for supervised scene prototyping, repetitive edits, material experiments, asset gathering, and workflows where visual feedback helps the model correct itself. Keep backups, review generated Python, split complex requests into smaller steps, and disable telemetry if its terms are unsuitable. It does not cover unattended background-mode Blender automation or provide an isolated execution sandbox.

Technical notes+

src/blender_mcp/server.py builds a FastMCP server around a persistent BlenderConnection, sends JSON over TCP, locks each send-and-receive cycle, detects complete responses by parsing accumulated bytes, and uses a 180-second timeout. addon.py listens on port 9876 by default, schedules received commands onto Blender's main thread with bpy.app.timers.register, and refuses Blender background mode. src/blender_mcp/telemetry.py queues events to a Supabase REST endpoint, persists an anonymous UUID, removes prompts and metadata without consent, and supports environment-based disabling. src/blender_mcp/telemetry_decorator.py can capture prompts, selected parameters, and generated code when consent permits. test_set_texture_version_guard.py is a static source-structure test for Blender shader-node compatibility rather than an end-to-end Blender test.

Observed

License
MIT
Primary language
Python
Packaging
Setuptools package with a PyPI command-line entry point; documented installation uses uvx or pipx, plus manual Blender add-on installation.
Interfaces
MCP server over standard client configuration, connected to the Blender add-on through a JSON TCP socket.
Runtime requirements
Python 3.10 or newer and Blender 3.0 or newer.
Platform support
Package metadata declares operating-system independence; setup instructions cover macOS, Windows, and Linux.
Testing structure
The supplied repository text includes one root-level static structural test for Blender version compatibility.

Read from README.md, pyproject.toml, main.py, src/blender_mcp/server.py, src/blender_mcp/__init__.py, src/blender_mcp/telemetry.py, src/blender_mcp/telemetry_decorator.py, addon.py, test_set_texture_version_guard.py, LICENSE, uv.lock, .python-version, TERMS_AND_CONDITIONS.md.

What it can do

  • Create 3D models from natural language descriptions

    Natural language prompt describing desired 3D model → 3D model in Blender

  • Modify existing 3D models through text commands

    Existing 3D model and natural language modification instructions → Modified 3D model

  • Generate 3D scenes from text descriptions

    Natural language description of desired scene → Complete 3D scene in Blender

  • Create and apply materials to 3D objects via prompts

    Natural language description of material properties → Material applied to 3D objects

  • Control Blender operations through conversational AI

    Natural language commands for Blender functions → Executed Blender operations

  • Manipulate 3D object properties via text instructions

    Text commands specifying object transformations or property changes → Updated 3D object properties in Blender

Tags

blender3d-modelingclaudemcpai-assistantautomation3d-designpython

Tech Stack

PythonDocker

Comments (0)

No comments yet

Editorially curated, with community endorsements as a secondary signal. Corrections welcome.