Skip to Content
Setup GuidesOverview

Blender MCP Setup Guides — Every Platform, Step by Step

Blender MCP connects AI to Blender 3D  through the Model Context Protocol. Choose your AI client below and follow the step-by-step guide to get started. Every guide covers prerequisites, installation, configuration, troubleshooting, and example prompts.

All guides require:

Choose Your AI Client

Listed from hardest to easiest, with the ready-to-use option last.

GuideAI ClientDifficultyCost
Ollama SetupLocal AI (Ollama)AdvancedNo hosted API bill — you supply and run a capable GPU
VSCode SetupVSCode + Cline / Roo CodeMediumFree
ChatGPT SetupChatGPT DesktopMediumChatGPT Plus recommended
Gemini SetupGoogle Gemini CLIMediumFree tier available
Claude SetupClaude Desktop / Claude CodeEasyFree tier available
Cursor SetupCursor IDEEasyFree tier available
3D-AgentIncluded — no MCP client neededEasiest — ready to use instantlyPaid plans, AI included

The setup guides connect an AI client you bring yourself; 3D-Agent is a paid app that includes the AI, so there is nothing to connect.

Claude is the recommended starting point — Blender MCP was designed for Claude, and it has the most mature integration. If you’re using Cursor for coding, the Cursor guide is your best bet.

Advanced: Self-Hosted Local Models

Want to run the model yourself instead of calling a hosted one? The Ollama guide covers it. You host the model and its local API on a capable GPU, and you configure and operate both that stack and MCP — substantially more setup than a hosted client, with results that are generally weaker than leading cloud models for complex Blender work.

How It Works

Every setup follows the same three-step pattern:

  1. Install the MCP server — The blender-mcp package (runs via uvx)
  2. Install the Blender addon — addon.py creates a socket server inside Blender
  3. Connect — Your AI client talks to the MCP server, which talks to Blender

For a technical deep dive into the architecture, see our MCP Server guide.

Last updated on
Claude