Blender MCP NextGen
Connects AI assistants to Blender 3D for natural language 3D modeling, scene inspection, material creation, and rendering.
README
Blender MCP NextGen 🍊⚡
Next-Generation Model Context Protocol (MCP) Server & Addon Suite — Antigravity & Claude Friendly!
Blender MCP NextGen connects AI assistants (such as Antigravity AI Agent, Claude Desktop, Cursor, VS Code, or custom LLMs) to Blender 3D using the official Model Context Protocol (MCP). It enables natural language 3D modeling, automated scene inspection, PBR material creation, viewport visual feedback, asset importing, and background rendering.
✨ Features & Enhancements
- 🤖 Antigravity & Claude Friendly: Native MCP stdio integration with ready-to-use config templates ([mcp_antigravity.json](file:///c:/Users/swatm/Downloads/Blender/mcp_antigravity.json) & [antigravity_instructions.md](file:///c:/Users/swatm/Downloads/Blender/antigravity_instructions.md)).
- 🛡️ AST Security Guardrails: Built-in Python AST analyzer inspects generated code before execution to prevent harmful system commands or unwanted operations.
- 🎨 Redesigned Blender Addon UI: A sleek, dark-themed N-Panel ("MCP Assistant") inside Blender featuring live server status badges, active TCP socket port settings, quick lighting tools, and a real-time execution console.
- 📸 Viewport Visual Feedback: Multi-modal AI support with
capture_viewportto take snapshot images of the active 3D viewport. - 🌿 PolyHaven Asset Library Search: Instantly query and import free HDRIs, 3D models, and PBR textures into your Blender scene.
- 💡 PBR Material Manager: Quick helper commands to apply custom metallic, roughness, and hex color properties to 3D meshes.
- 🌐 Web Setup Dashboard: Interactive local dashboard for generating MCP client JSON configurations and viewing server diagnostics.
🚀 Quick Start Guide
1. Requirements
- Blender: 3.0 or higher (compatible with Blender 4.x)
- Python: 3.10 or higher
2. Install Server Dependencies
pip install -r requirements.txt
3. Install Addon in Blender
- Open Blender.
- Go to Edit > Preferences > Add-ons.
- Click Install... at the top right and select
addon/blender_mcp_addon.py. - Check the box to enable Blender MCP NextGen.
- In the 3D Viewport, press N to expand the sidebar and click the MCP Assistant tab.
- Click Start Listener (Status should change to ONLINE 🟢).
⚙️ MCP Client Configurations
For Antigravity AI Agent (mcp_antigravity.json)
{
"mcpServers": {
"blender-nextgen": {
"command": "python",
"args": ["c:/Users/swatm/Downloads/Blender/server/blender_mcp_server.py"]
}
}
}
For Claude Desktop (claude_desktop_config.json)
{
"mcpServers": {
"blender-nextgen": {
"command": "python",
"args": ["c:/Users/swatm/Downloads/Blender/server/blender_mcp_server.py"]
}
}
}
For Cursor IDE (.cursor/mcp.json)
{
"mcpServers": {
"blender-nextgen": {
"command": "python",
"args": ["c:/Users/swatm/Downloads/Blender/server/blender_mcp_server.py"]
}
}
}
🛠️ MCP Tools Reference
| Tool Name | Description |
|---|---|
execute_blender_code |
Executes safe Python code inside Blender with stdout, stderr, and error tracebacks. |
get_scene_info |
Returns structured JSON topology of objects, lights, cameras, active materials, and render settings. |
capture_viewport |
Takes a screenshot snapshot of the current 3D viewport for visual AI context. |
fetch_polyhaven_asset |
Downloads and imports HDRIs, textures, or models from PolyHaven. |
manage_material |
Creates or updates PBR materials with hex colors, metallic, and roughness values. |
render_scene |
Triggers Eevee or Cycles rendering and saves the output. |
🔒 Security Notice
All code sent via execute_blender_code is parsed and checked against server/security_guardrails.py. Forbidden modules (os, sys, subprocess, shutil) and execution statements (eval, exec on system level) are blocked by default.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。