Blender + MCP: how AI agents drive 3D software
13 steps · about 6 h 3 min · Intermediate
- Introducing the Model Context Protocol (article, Anthropic, 8 min)
The original announcement explains what problem MCP solves and the server/client split that every Blender add-on builds on.
- What is MCP? Explained in 4 Minutes (video, YouTube, 5 min)
A short visual recap of how AI apps use MCP to reach tools and data, before the technical reading.
- Model Context Protocol (article, Wikipedia, 10 min)
Gives the vocabulary of hosts, clients, servers, tools and resources that the Blender repos use.
- Build an MCP server (tutorial, modelcontextprotocol.io, 45 min)
Building a small tool-exposing server yourself shows what a Blender MCP server does, just with weather instead of meshes.
- MCP for Beginners (repo, Microsoft (GitHub), 1 h 30 min)
A free curriculum with first-server and first-client modules if you want more depth on transports and protocol features.
- Scripting Introduction - Blender Manual (reference, Blender Foundation, 10 min)
Introduces Blender's Python console, text editor and the official starting points for scripting.
- Quickstart - Blender Python API (doc, Blender Foundation, 30 min)
Covers bpy.data, context and operators, the exact calls an AI agent makes when it edits your scene.
- Getting Started with Scripting in Python (BCON23) (video, Mike Shah, Blender Conference, 40 min)
A conference talk that walks beginners through writing and packaging their first Blender add-on, which is the other half of every MCP integration.
- 3D Programming for Beginners Using Python and Blender 2.8, Tutorial (video, Olav3D Tutorials, 30 min)
A hands-on first script that creates geometry from code, the same skill the AI uses when it generates Blender Python.
- Blender Python API Documentation (doc, Blender Foundation, 20 min)
Shows the Gotchas and API Overview sections, including why threads are unsupported, which explains the design of MCP add-ons.
- MCP for Blender (repo, Siddharth Ahuja (GitHub), 30 min)
The widely used original: read how the add-on's socket server and the Python MCP server connect an LLM to Blender.
- I Gave AI Access to Blender. Here's What It Built (article, DEV Community, 15 min)
A practical walkthrough of installing the add-on and prompting a scene, so you see the full loop in action.
- Blender_mcp: Control Blender with AI via the Model Context Protocol (repo, glonorce (GitHub), 30 min)
A more advanced variant that shows thread-safe bpy execution and a stdio bridge, which shows how the design scales to many tools.