Godot AI
An asset by Dsarno
Install Asset
Install via Godot
To maintain one source of truth, Godot Asset Library is just a mirror of the old asset library so you can download directly on Godot via the integrated asset library browser
Quick Information
Godot AI
Godot AI connects MCP-compatible AI assistants (Claude Code, Codex, Antigravity, etc.) directly to a live Godot editor via the Model Context Protocol. Inspect, build, paint, and animate your scenes from a prompt — create nodes, wire signals, compose UI, author materials and particle effects, run tests, and read editor data without leaving chat. The third preview image was built from two prompts in five minutes — no manual editor work.From the creators of MCP for Unity — the most widely used gamedev MCP, with 8,500+ GitHub stars and hundreds of thousands of users — built from the ground up for Godot. Robustness and deep QA are core to our approach, shaped by over a year of development and community feedback on MCP for Unity.Features:- 80+ MCP tools spanning scenes, nodes, scripts, animations, UI, themes, materials, particles, input mapping, and project settings- Materials: StandardMaterial3D, ORMMaterial3D, and ShaderMaterial — plus six presets (metal, glass, emissive, unlit, matte, ceramic)- Particles: GPU + CPU, 2D + 3D, with seven presets (fire, smoke, spark_burst, magic, rain, explosion, lightning) — one call spawns emitter, process material, mesh, and billboard as a single undoable action- Smart screenshots so AI agents can see the scene- Build entire UI hierarchies from a single prompt- One-click client setup for Claude Code, Codex, and Antigravity- Multi-instance support across multiple Godot editors- Auto-managed Python server — enable the plugin and goRequirements: Godot 4.3+ (4.4+ recommended) and an MCP client.Quick Start:1. Install uv (one-time): macOS/Linux: curl -LsSf https://astral.sh/uv/install.sh | sh Windows (PowerShell): irm https://astral.sh/uv/install.ps1 | iex2. Install the addon from the Asset Library (or drop plugin/addons/godot_ai into your project's addons/ folder).3. In Godot: Project > Project Settings > Plugins — enable "Godot AI".4. Open the Godot AI dock and click your MCP client to configure it in one click.5. Open your MCP client and start prompting.Full docs: https://github.com/hi-godot/godot-aiComing next: Physics, Shaders, Terrain, Custom Tools, More Client Configurations
Supported Engine Version
4.3
Version String
0.4.3
License Version
MIT
Support Level
community
Modified Date
10 hours ago
Git URL
Issue URL
Godot AI
Connect AI assistants to a live Godot editor via the Model Context Protocol (MCP).
Godot AI bridges Claude Code, Codex, Antigravity, and other MCP clients with your editor — inspect scenes, create nodes, modify properties, run tests, search project files, and more, all from a prompt.
Quick Start
- Copy
addons/godot_ai/into your project'saddons/folder - Enable the plugin: Project > Project Settings > Plugins > Godot AI
- Pick your MCP client in the Godot AI dock and press Configure
The plugin auto-starts the MCP server and connects over WebSocket. No manual configuration required.
Requirements
- Godot 4.3+ (4.4+ recommended)
- uv (used to install the Python server)
- An MCP client (Claude Code | Codex | Antigravity)
Documentation
Full documentation, contributing guide, and source code: github.com/hi-godot/godot-ai
License
Godot AI connects MCP-compatible AI assistants (Claude Code, Codex, Antigravity, etc.) directly to a live Godot editor via the Model Context Protocol. Inspect, build, paint, and animate your scenes from a prompt — create nodes, wire signals, compose UI, author materials and particle effects, run tests, and read editor data without leaving chat. The third preview image was built from two prompts in five minutes — no manual editor work.
From the creators of MCP for Unity — the most widely used gamedev MCP, with 8,500+ GitHub stars and hundreds of thousands of users — built from the ground up for Godot. Robustness and deep QA are core to our approach, shaped by over a year of development and community feedback on MCP for Unity.
Features:
- 80+ MCP tools spanning scenes, nodes, scripts, animations, UI, themes, materials, particles, input mapping, and project settings
- Materials: StandardMaterial3D, ORMMaterial3D, and ShaderMaterial — plus six presets (metal, glass, emissive, unlit, matte, ceramic)
- Particles: GPU + CPU, 2D + 3D, with seven presets (fire, smoke, spark_burst, magic, rain, explosion, lightning) — one call spawns emitter, process material, mesh, and billboard as a single undoable action
- Smart screenshots so AI agents can see the scene
- Build entire UI hierarchies from a single prompt
- One-click client setup for Claude Code, Codex, and Antigravity
- Multi-instance support across multiple Godot editors
- Auto-managed Python server — enable the plugin and go
Requirements: Godot 4.3+ (4.4+ recommended) and an MCP client.
Quick Start:
1. Install uv (one-time):
macOS/Linux: curl -LsSf https://astral.sh/uv/install.sh | sh
Windows (PowerShell): irm https://astral.sh/uv/install.ps1 | iex
2. Install the addon from the Asset Library (or drop plugin/addons/godot_ai into your project's addons/ folder).
3. In Godot: Project > Project Settings > Plugins — enable "Godot AI".
4. Open the Godot AI dock and click your MCP client to configure it in one click.
5. Open your MCP client and start prompting.
Full docs: https://github.com/hi-godot/godot-ai
Coming next: Physics, Shaders, Terrain, Custom Tools, More Client Configurations
Reviews
Quick Information
Godot AI
Godot AI connects MCP-compatible AI assistants (Claude Code, Codex, Antigravity, etc.) directly to a live Godot editor via the Model Context Protocol. Inspect, build, paint, and animate your scenes from a prompt — create nodes, wire signals, compose UI, author materials and particle effects, run tests, and read editor data without leaving chat. The third preview image was built from two prompts in five minutes — no manual editor work.From the creators of MCP for Unity — the most widely used gamedev MCP, with 8,500+ GitHub stars and hundreds of thousands of users — built from the ground up for Godot. Robustness and deep QA are core to our approach, shaped by over a year of development and community feedback on MCP for Unity.Features:- 80+ MCP tools spanning scenes, nodes, scripts, animations, UI, themes, materials, particles, input mapping, and project settings- Materials: StandardMaterial3D, ORMMaterial3D, and ShaderMaterial — plus six presets (metal, glass, emissive, unlit, matte, ceramic)- Particles: GPU + CPU, 2D + 3D, with seven presets (fire, smoke, spark_burst, magic, rain, explosion, lightning) — one call spawns emitter, process material, mesh, and billboard as a single undoable action- Smart screenshots so AI agents can see the scene- Build entire UI hierarchies from a single prompt- One-click client setup for Claude Code, Codex, and Antigravity- Multi-instance support across multiple Godot editors- Auto-managed Python server — enable the plugin and goRequirements: Godot 4.3+ (4.4+ recommended) and an MCP client.Quick Start:1. Install uv (one-time): macOS/Linux: curl -LsSf https://astral.sh/uv/install.sh | sh Windows (PowerShell): irm https://astral.sh/uv/install.ps1 | iex2. Install the addon from the Asset Library (or drop plugin/addons/godot_ai into your project's addons/ folder).3. In Godot: Project > Project Settings > Plugins — enable "Godot AI".4. Open the Godot AI dock and click your MCP client to configure it in one click.5. Open your MCP client and start prompting.Full docs: https://github.com/hi-godot/godot-aiComing next: Physics, Shaders, Terrain, Custom Tools, More Client Configurations
Supported Engine Version
4.3
Version String
0.4.3
License Version
MIT
Support Level
community
Modified Date
10 hours ago
Git URL
Issue URL