dcc-mcp-godot
Enables Godot 4 editor interaction through MCP, including project/scene inspection, script writing, and a 2D roguelike skill for AI agents.
README
dcc-mcp-godot
Godot 4 editor adapter for the DCC Model Context Protocol ecosystem. It ships a GDScript
EditorPlugin, typed project and scene tools, and a tested 2D arena-roguelike skill.
Install
pip install dcc-mcp-godot
Install the packaged addon into a Godot project:
dcc-mcp-godot-install /path/to/project
Enable DCC-MCP Godot under Project Settings > Plugins, then run:
dcc-mcp-godot
The MCP endpoint defaults to http://127.0.0.1:8765/mcp. The plugin connects to the
loopback bridge at ws://127.0.0.1:3847; override the port with
DCC_MCP_GODOT_BRIDGE_PORT before starting both processes.
Agent workflow
- Load
godot-projectto inspect the project or write a bounded.gdfile belowres://. - Load
godot-sceneto inspect the current scene and make undoable node edits. - Load
godot-roguelike, callcreate_2d_roguelike, thenvalidate_2d_roguelike. - Run the project. Arrow keys move the player; enemies spawn continuously and the player auto-attacks, earns experience, levels up, and eventually reaches game over.
No raw editor evaluation or arbitrary filesystem command is exposed. Scene edits use Godot's
undo manager; script writes are restricted to .gd files below res://.
Real Godot CI
CI resolves the official godotengine/godot latest stable GitHub release and downloads the Linux
editor. It starts the packaged MCP server and EditorPlugin, sends real JSON-RPC load_skill,
paginated tools/list, tools/call, and jobs_get_status requests, then runs the generated game
simulation and requires ROGUELIKE_SMOKE_OK.
Run the same smoke locally:
python tools/download_latest_godot.py --output .godot-bin
python tests/live_godot_smoke.py --godot /path/to/Godot
Development
python -m pip install -e ".[dev]"
pytest
ruff check src tests tools
ruff format --check src tests tools
python tools/lint_skills.py
python -m build
python -m twine check dist/*
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。