Blockbench MCP

https://github.com/user-attachments/assets/ab1b7e63-b6f0-4d5b-85ab-79d328de31db

Plugin Installation

Open the desktop version of Blockbench, go to File > Plugins and click the "Load Plugin from URL" and paste in this URL:

https://achmadawdi.github.io/mcp-blockbench/mcp.js

Model Context Protocol Server

Configure the MCP server under Blockbench settings: Settings > General > MCP Server Port and MCP Server Endpoint

The following examples use the default values of :3000/bb-mcp

Multiple Blockbench Windows

When more than one Blockbench window is open, the plugin can auto-select the next available port (Settings > General > Auto-select Available Port, enabled by default).

Example:

• Window 1: http://localhost:3000/bb-mcp
• Window 2: http://localhost:3001/bb-mcp
• Window 3: http://localhost:3002/bb-mcp

Use the MCP panel in each Blockbench window to copy the exact URL or client config snippet. Each window needs a unique mcpServers key in your client config.

Cursor — two projects:

{
  "mcpServers": {
    "blockbench_project_a": {
      "url": "http://localhost:3000/bb-mcp"
    },
    "blockbench_project_b": {
      "url": "http://localhost:3001/bb-mcp"
    }
  }
}

Codex — two projects:

[mcp_servers.blockbench_project_a]
url = "http://localhost:3000/bb-mcp"

[mcp_servers.blockbench_project_b]
url = "http://localhost:3001/bb-mcp"

Settings apply on plugin reload or Blockbench restart.

Per-Face UV for Texture Painting

For custom texture atlases (furniture, items, multi-part models), use per-face UV:

1. create_project with box_uv: false and set texture_width / texture_height
2. place_cube with explicit { face, uv } rectangles
3. get_uv_layout to export face regions before painting
4. Use the texture_uv_workflow MCP prompt for the full guide

Installation

General

npx mcp-add --type http --url "http://localhost:3000/bb-mcp" --scope project

VS Code

.vscode/mcp.json

{
  "servers": {
    "blockbench": {
      "url": "http://localhost:3000/bb-mcp",
      "type": "http"
    }
  }
}

Claude Desktop

claude_desktop_config.json

{
  "mcpServers": {
    "blockbench": {
      "command": "npx",
      "args": ["mcp-remote", "http://localhost:3000/bb-mcp"]
    }
  }
}

Claude Code

claude mcp add blockbench --transport http http://localhost:3000/bb-mcp

Cursor

.cursor/mcp.json (project) or ~/.cursor/mcp.json (global)

{
  "mcpServers": {
    "blockbench": {
      "url": "http://localhost:3000/bb-mcp"
    }
  }
}

Blockbench must be running with the plugin loaded. Enable the server in Settings > MCP.

Codex

~/.codex/config.toml (or project .codex/config.toml in trusted projects)

Direct Streamable HTTP:

[mcp_servers.blockbench]
url = "http://localhost:3000/bb-mcp"

If direct HTTP fails to list tools (known Codex CLI limitations), use the stdio bridge:

[mcp_servers.blockbench]
command = "npx"
args = ["mcp-remote", "http://localhost:3000/bb-mcp"]

Antigravity

{
  "mcpServers": {
    "blockbench": {
      "serverUrl": "http://localhost:3000/bb-mcp"
    }
  }
}

Cline

<img width="674" height="486" alt="Connecting to Blockbench MCP plugin through Cline" src="https://github.com/user-attachments/assets/f27f2304-dd56-4c60-b159-86fbd5af65ee" />

cline_mcp_settings.json

{
  "mcpServers": {
    "blockbench": {
      "url": "http://localhost:3000/bb-mcp",
      "type": "streamableHttp",
      "disabled": false,
      "autoApprove": []
    }
  }
}

Ollama

uvx ollmcp -u http://localhost:3000/bb-mcp

Recommended: jonigl/mcp-client-for-ollama

OpenCode

opencode mcp add

<img width="504" height="300" alt="Connecting to Blockbench MCP plugin through OpenCode." src="https://github.com/user-attachments/assets/238971fc-0048-4b8d-95dd-6681604bbe90" />

Usage

See the docs for tool and prompt reference.

Use Agent Skills in .agents/skills/ to orchestrate tool usage.

Attribution

Forked from blockbench-mcp-plugin by Jason J. Gardner (GPL-3.0). Maintained independently at achmadawdi/mcp-blockbench.

Plugin Development

See CONTRIBUTING.md for detailed instructions on setting up the development environment and how to add new tools, resources, and prompts.