Godot MCP Server

Community

MCP server for interfacing with the Godot game engine

How to install ↓ Compare GitHub source repository ↗

Category: Dev Tools #108 of 438
Stars: ★ 6.1k Popular
Transport: stdio (local process)
Runtime: Node.js 18+ · Prebuilt binary
Credentials: No credential needed
License: MIT
Last commit: 6 months ago ⚠
Tools: 14

61FMRS · C

This server provides a bridge for AI agents to interact with the Godot engine, supporting editor launching, project running, debug output capture, and more. Its GDScript integration allows complex scene operations, making it suitable for AI-assisted game development. Installation is straightforward and compatible with multiple MCP clients.

Strongest · Setup experience 14/20 Weakest · Reliability 9/20

Reliability

9/20

Security and permissions

11/20

Maintenance

14/20

Documentation

13/20

Setup experience

14/20

Why each score

Reliability 9/20: The server source code is structurally complete, with a clear TypeScript build and main entry. Tool declarations appear consistent with implementations. However, no tests or CI workflows are provided, so reproducibility of key paths cannot be confirmed from static evidence. Error handling and edge cases (e.g., Godot not installed, invalid paths) are only partially handled in code, and the troubleshooting docs do not fully align with implementation. Therefore reliability is scored 9.
Security and permissions 11/20: The server invokes Godot via command line and has no credential handling. Environment variables GODOT_PATH and DEBUG are reasonable. There is no external network exfiltration or malicious operations. However, there is no explicit confirmation mechanism for dangerous operations (e.g., stopping projects, creating scenes), and the tool list lacks permission tiers or user confirmation steps. The Cline example suggests auto-approving all tools, but the server itself does not enforce least privilege or confirmation. No red-line issues found, so score is 11.
Maintenance 14/20: The repository is active with 5104 stars and recent commits. MIT license is present, and maintainer provides sponsorship links. However, there is no evidence of versioned releases, dependency update policy, or issue response patterns. Therefore, maintenance is scored 14.
Documentation 13/20: The README provides detailed installation instructions, environment variable descriptions, client configuration examples (Claude, Cline, Cursor), and troubleshooting. However, it lacks detailed tool parameter documentation, usage limits, cost information, and test evidence. Documentation generally matches capabilities but contains hidden assumptions (e.g., Godot path detection). Therefore, documentation is scored 13.
Setup experience 14/20: Installation steps are clear, supporting npx and building from source, with configuration examples for multiple MCP clients. However, it depends on external Godot engine installation, lacks platform-specific instructions for Windows/macOS, and source build requires manual steps. Due to missing tests and CI, setup is capped at 15 and scored 14 based on static evidence.

Static review · not runListed 2026-08-07

Read the FMRS scoring method →

Fit and risk

What it can accessReads local filesWrites / deletes local filesRuns commands or code

Best for

AI agents working with the Godot engine for development assistance
Developers who want to automate project launching and debug feedback loops
AI tools that need to dynamically create and modify scenes in Godot projects

Not for

Developers not using the Godot engine
Users requiring remote or HTTP-based connections (only stdio transport)
Users without a need for Godot 4.4+ UID features

Required permissions

Execute the Godot executable (to launch editor, run projects)
Read specified project directories to list projects and get information
Write files in the project directory to create scenes, add nodes, and save scenes
Read console output to capture debug information

Risks and side effects

Running untrusted Godot projects may execute arbitrary code, posing security risks
Automatic modification of project files may lead to unintended changes; version control is recommended
Operations may fail if the Godot version or path is configured incorrectly
High volumes of operations may consume system resources, affecting performance

Setup

Before you start

Runtime：Node.js 18+ · Prebuilt binary

Other optional settings (2)

GODOT_PATH optional Path to the Godot executable, overriding automatic detection; set it to the binary from your install at godotengine.org. Only needed if auto-detection fails.

DEBUG optional Set to "true" to enable detailed server-side debug logging; optional, not a secret.

Ensure the Godot Engine is installed on your system.\n2. Ensure Node.js (>=18.0.0) and npm are installed.\n3. For Claude Code, run: claude mcp add godot -- npx @coding-solo/godot-mcp.\n4. For Cline, add the provided configuration to the Cline MCP settings file.\n5. For Cursor, add the server via the UI or a project-specific mcp.json.\n6. Optionally, set environment variables GODOT_PATH and DEBUG before running.

claude_desktop_config.json

{
  "mcpServers": {
    "godot": {
      "command": "npx",
      "args": [
        "@coding-solo/godot-mcp"
      ],
      "env": {
        "GODOT_PATH": "/path/to/godot",
        "DEBUG": "true"
      }
    }
  }
}

Shown for Claude Desktop. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.

.vscode/mcp.json

{
  "servers": {
    "godot": {
      "command": "npx",
      "args": [
        "@coding-solo/godot-mcp"
      ],
      "env": {
        "GODOT_PATH": "/path/to/godot",
        "DEBUG": "true"
      }
    }
  }
}

Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).

Terminal

claude mcp add godot -e GODOT_PATH=/path/to/godot -e DEBUG=true -- npx @coding-solo/godot-mcp

Run it in a terminal; replace any <…> placeholders with your own values first.

Generate and validate client config

Check that it works

After restarting the client, 14 tools including launch_editor, run_project, and get_debug_output should appear in the tool list; alternatively ask "What Godot version is installed?" — a version returned by get_godot_version confirms the connection.

Troubleshooting

If Godot is not found, set the environment variable `GODOT_PATH` to your Godot executable path.
If connection issues occur, ensure the server is running and restart your AI assistant.
If the project path is invalid, ensure the path points to a directory containing a `project.godot` file.
If build issues arise, run `npm install` to install all dependencies.
For Cursor users, ensure the MCP server is enabled in settings and use the Agent chat profile (requires Cursor Pro or Business subscription).

Things to try

Once connected, you can ask your AI assistant things like:

Launch the Godot editor for my current project
Run this Godot project in debug mode and show me the debug output
List all Godot projects in the ~/games directory
Create a new scene with a Node2D root node and add a Sprite2D node to it

Tools 14

launch_editor writes

Launch the Godot editor for a specific project.

run_project writes

Run a Godot project in debug mode.

get_debug_output read-only

Retrieve console output and error messages.

stop_project writes

Stop a running Godot project programmatically.

get_godot_version read-only

Retrieve the installed Godot version.

list_projects read-only

Find Godot projects in a specified directory.

get_project_info read-only

Get detailed information about the project structure.

create_scene writes

Create a new scene with a specified root node type.

Show 6 more tools

add_node writes

Add nodes to an existing scene with customizable properties.

load_sprite writes

Load sprites and textures into Sprite2D nodes.

export_mesh_library writes

Export 3D scenes as MeshLibrary resources for GridMap.

save_scene writes

Save scenes with options for creating variants.

get_uid read-only

Get the UID for a specific file (Godot 4.4+).

update_project_uids writes

Update UID references by resaving resources (Godot 4.4+).

Use cases

AI-assisted development and debugging of Godot projects

Automated scene creation and node manipulation

Real-time capture of game run output during coding sessions

Batch management of projects and scene resources

Supported clients

Claude Desktop

Cline

Cursor

Listed from the project's documentation, not tested by this site.

Overview

The Godot MCP server enables AI agents to launch the Godot editor, run projects, capture debug output, and control project execution. It provides a direct feedback loop that helps agents understand what works and what doesn't in real Godot projects, leading to better code generation and debugging assistance. The server uses a bundled GDScript script to handle complex operations such as creating scenes, adding nodes, loading sprites, and exporting MeshLibrary resources.

Similar servers

Context7 80 · B

Upstash's official server providing up-to-date third-party library docs for AI coding assistants

★ 63.3k · Tools 2 Compare with this →

NPM Sentinel MCP Server 79 · B

AI-powered NPM package analysis MCP server

★ 18 · Tools 19 Compare with this →

MartinLoop MCP Server 79 · B

An independent control layer that gives AI coding agents budgets, brakes, and receipts

★ 413 Compare with this →

GitHub MCP Server 76 · B

GitHub's official MCP server for managing repos, issues, PRs, and workflows via natural language

★ 33.7k · Tools 15 Compare with this →

Compare with other MCP servers

Source revision 1209744fad78 Data synced 2026-10-19 Read the FMRS scoring method