A collection of demo and template projects for the Godot Engine, covering 2D and 3D game examples written in GDScript.
limbonaut/limboai
A C++ plugin for Godot 4 that adds behavior trees and hierarchical state machines for building game AI, with an editor and visual debugger.
- Stars
- 3,074
- Forks
- 150
- Open issues
- 62
- Last commit
- 18 days ago
What is it?
LimboAI is a plugin for Godot Engine 4 that provides behavior trees and state machines, which can be combined to build complex agent behaviors such as characters and enemies. It includes a behavior tree editor, built-in class documentation, a visual debugger that inspects running trees, and a demo project with a tutorial.
Behavior trees are built from composite, decorator, condition and action tasks, and custom tasks can be written by extending the core classes in C++ or GDScript. A blackboard system shares data between tasks, with plans, scopes and exported parameters for reuse across agents. Hierarchical state machines use event-based transitions and can be nested, with BTState bridging the two systems.
The project can be used as a C++ engine module or as a GDExtension shared library, so custom engine builds are not required. State machines are configured in code, since there is no GUI editor for them.
Key features
- 01Behavior tree editor for creating and saving BehaviorTree resources
- 02Visual debugger for inspecting running behavior trees
- 03Blackboard system with plans, scopes and exported parameters
- 04Hierarchical state machines with event-based transitions
- 05Custom tasks in C++ or GDScript
Use cases
- Designing enemy or NPC AI in a Godot 4 game
- Prototyping state-driven character logic quickly in a game jam
- Organizing reusable AI behaviors across several agents
Quick start
- 1
Download a precompiled build or the demo project from the Releases page
- 2
Open demo/scenes/showcase.tscn in Godot to see the examples
- 3
To build the GDExtension from source, you need SCons and a C++ compiler, then run the build
scons target=editor
Things to know
- The GDExtension version requires Godot 4.6 or higher and has somewhat fewer features than the C++ module.
- State machine setup and initialization require code, since there is no GUI editor for them.
Topics
FAQ
- What is limboai?
- A C++ plugin for Godot 4 that adds behavior trees and hierarchical state machines for building game AI, with an editor and visual debugger. LimboAI is a plugin for Godot Engine 4 that provides behavior trees and state machines, which can be combined to build complex agent behaviors such as characters and enemies. It includes a behavior tree editor, built-in class documentation, a visual debugger that inspects running trees, and a demo project with a tutorial.
- Is limboai free to use? What license does it have?
- limboai is open source under the MIT license. Check the license terms before using it in commercial products.
- Is limboai still maintained?
- Actively maintained. The last commit to limboai was on Oct 4, 2026.
- What can I use limboai for?
- Designing enemy or NPC AI in a Godot 4 game Prototyping state-driven character logic quickly in a game jam Organizing reusable AI behaviors across several agents
- How do I install limboai?
- Download a precompiled build or the demo project from the Releases page Open demo/scenes/showcase.tscn in Godot to see the examples To build the GDExtension from source, you need SCons and a C++ compiler, then run the build: scons target=editor
- Where can I get limboai?
- The source code and installation instructions are on GitHub: https://github.com/limbonaut/limboai
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A standalone JavaScript library for game AI, covering steering, navigation meshes, perception, fuzzy logic and state-driven agents.
A Python library for building decision-making applications like chatbots and agents as state machines, with a built-in UI for tracing and monitoring.
Unity ML-Agents Toolkit lets Unity games and simulations serve as environments for training AI agents with reinforcement and imitation learning.
Open-source interactive world model that renders playable game scenes from low-poly frames using streaming neural rendering.
Open-source interactive game video generation model that turns a reference image and keyboard or mouse inputs into controllable gameplay-style video.