godot-mcp

godot-mcp

MCP bridge between an AI assistant and the Godot editor, enabling control via 6 tools (call methods, run GDScript, screenshots, etc.) over WebSocket.

Category
Visit Server

README

🇫🇷 Français · 🇬🇧 English


<a id="fr"></a>godot-mcp

Pont MCP entre un assistant IA et l'éditeur Godot.

Assistant IA <---stdio/MCP---> godot-mcp <---WebSocket:6505---> Plugin Godot

6 outils MCP pour piloter l'éditeur : godot_call (175+ méthodes), godot_list_methods, godot_info, godot_screenshot, godot_execute, godot_status.

Installation

1. Plugin Godot

Copiez le dossier plugin/ dans addons/godot_mcp/ de votre projet Godot.

votre-projet/
└── addons/
    └── godot_mcp/       ← copiez plugin/ ici

Activez-le : Projet → Paramètres du projet → Plugins → godot-mcp → Activer

2. Serveur

npm install && npm run build && npm start

3. Client MCP

Ajoutez à la config de votre client IA (.mcp.json, claude.json, opencode.json) :

{
  "mcpServers": {
    "godot-mcp": {
      "command": "node",
      "args": ["/chemin/vers/godot-mcp/dist/index.js"]
    }
  }
}

Outils

Outil Description
godot_call Appelle n'importe quelle méthode (175+)
godot_list_methods Liste les méthodes par catégorie
godot_info Infos projet
godot_screenshot Capture éditeur en PNG
godot_execute Exécute du GDScript
godot_status Vérifie la connexion

Chemins de nodes : les paramètres parent_path et node_path de godot_call (ex. add_node, update_property, delete_node) sont toujours relatifs à la racine de la scène actuellement éditée ("." = racine de la scène). Les chemins absolus Godot ("/root", "../...") sont rejetés — ils cibleraient l'arbre interne de l'éditeur au lieu de la scène.

godot_screenshot nécessite un éditeur avec rendu actif : il ne fonctionne pas en mode --headless (erreur "Could not get image from viewport").

Arborescence

godot-mcp/
├── plugin/              ← Plugin Godot (à copier dans votre projet)
├── src/index.ts         ← Serveur MCP (Node.js)
├── package.json
├── tsconfig.json
├── README.md
└── LICENSE              ← MIT

Licence

MIT

Ce plugin Godot dérive de godot-mcp-pro (Youichi Uda, MIT).


<a id="en"></a>godot-mcp

MCP bridge between an AI assistant and the Godot editor.

AI Assistant <---stdio/MCP---> godot-mcp <---WebSocket:6505---> Godot Plugin

6 MCP tools to control the editor: godot_call (175+ methods), godot_list_methods, godot_info, godot_screenshot, godot_execute, godot_status.

Setup

1. Godot Plugin

Copy the plugin/ folder into your project's addons/godot_mcp/.

your-project/
└── addons/
    └── godot_mcp/       ← copy plugin/ here

Enable it: Project → Project Settings → Plugins → godot-mcp → Enable

2. Server

npm install && npm run build && npm start

3. MCP Client

Add to your AI client config (.mcp.json, claude.json, opencode.json):

{
  "mcpServers": {
    "godot-mcp": {
      "command": "node",
      "args": ["/path/to/godot-mcp/dist/index.js"]
    }
  }
}

Tools

Tool Description
godot_call Call any method (175+)
godot_list_methods List methods by category
godot_info Project info
godot_screenshot Editor screenshot in PNG
godot_execute Run GDScript
godot_status Check connection

Node paths: parent_path and node_path parameters of godot_call (e.g. add_node, update_property, delete_node) are always relative to the root of the currently edited scene ("." = scene root). Absolute Godot paths ("/root", "../...") are rejected — they would target the editor's internal tree instead of the scene.

godot_screenshot requires an editor with active rendering: it does not work in --headless mode ("Could not get image from viewport" error).

Structure

godot-mcp/
├── plugin/              ← Godot plugin (copy to your project)
├── src/index.ts         ← MCP server (Node.js)
├── package.json
├── tsconfig.json
├── README.md
└── LICENSE              ← MIT

License

MIT

This Godot plugin is derived from godot-mcp-pro (Youichi Uda, MIT).

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured