Ecosystem

One app. Three engines. 27 extensions.

AI Game Developer is modular: the desktop app and the gamedev CLI drive the plugins for Unity, Unreal Engine and Godot, each plugin has its own CLI and extensions, and the bridge underneath is open-source .NET. Install only what you need.

3
engines
27
extensions
14
AI clients

How it fits together

  1. Get the app — or a CLI

    The desktop app is the easiest start; the gamedev CLI and the engine CLIs do the same from a terminal.

    Download →
  2. Add the plugin to your engine

    The app installs the Unity, Unreal Engine or Godot plugin into your project and signs it in.

    Install guides →
  3. Add extensions you need

    Animation, terrain, cameras, navigation… each extension adds one area of the engine.

    Browse extensions →
  4. Use any AI client

    Chat in the app, or point Claude, Cursor, Copilot, Gemini and others at the MCP endpoint.

    MCP endpoint →

The app

One place to build, edit and playtest your game with AI. It installs the engine plugins for you.

AI Game Dev

The desktop app: an AI agent that builds, edits and playtests your game live inside Unity, Unreal Engine and Godot.

  • Sets up the engine plugin in your project for you
  • Packages add Blender 3D modeling, local image and music models and more
  • Works with your ai-game.dev account; Windows, macOS and Linux
  • Updates itself

gamedev CLI

The app's agent in your terminal, with the same account and projects.

  • Install with one npm command
  • Sets up engine projects and talks to the same agent
  • Windows, macOS and Linux
npm install -g @baizor/gamedev

Engine plugins

The part inside your editor: it gives the AI agent the engine's tools over MCP.

Unity MCP

The Unity Editor plugin: AI drives scenes, GameObjects, assets, prefabs and scripts.

  • MCP tools plus RQL, one query language for the whole project
  • Unity 2022.3 LTS and Unity 6
  • Works in the Editor and inside built games
  • Add your own tools with C# attributes
openupm add com.ivanmurzak.unity.mcp

Unreal MCP

The Unreal Engine plugin: AI works with levels, actors, Blueprints and assets.

  • Unreal Engine 5.5 and newer, C++ projects
  • RQL queries, Blueprint, source and Python tools
  • Windows, macOS and Linux
  • Extensions for Niagara, PCG, GAS and more
unreal-mcp-cli install-plugin ./YourProject

Godot MCP

The Godot addon: AI edits scenes, nodes, resources and C# scripts.

  • Godot 4.3 and newer, .NET build
  • MCP tools plus RQL for scenes and resources
  • Optional tools inside your running game
  • Extensions for Beehave, Dialogic, Terrain3D and more
godot-cli install-plugin ./MyGodotProject

Extensions

Optional packages that teach the agent a whole area of an engine — 27 across three engines.

Engine CLIs

Set up a project from a terminal or CI: install the plugin, write your AI client's config, run tools.

unity-mcp-cli

Unity from the terminal: create projects, install the Editor and the plugin, configure AI clients.

  • Install a Unity Editor and create a project
  • Write the MCP config for 14 AI clients
  • Run tools and wait for the Editor in CI
npm install -g unity-mcp-cli

unreal-mcp-cli

Unreal Engine from the terminal: install the plugin, sign in, configure AI clients.

  • Install or remove the plugin and extensions
  • Write the MCP config for your AI client
  • Open the editor, wait for it, run tools
npm install -g unreal-mcp-cli

godot-cli

Godot from the terminal: install the addon and its packages, configure AI clients.

  • Install the addon, NuGet packages and extensions
  • Write the MCP config for your AI client
  • Build, open and wait for the editor
npm install -g godot-cli

.NET libraries

The open-source bridge under every plugin, reusable in your own .NET app.

MCP Plugin for .NET

The bridge between an app and an MCP server: expose .NET methods as MCP tools.

  • Tools, prompts and resources from C# attributes
  • Real-time connection to the MCP server
  • Powers the Unity, Unreal and Godot plugins
dotnet add package com.IvanMurzak.McpPlugin

ReflectorNet

Reflection for AI: serialize, inspect and change .NET objects from JSON.

  • JSON schemas for your types and methods
  • Find methods by approximate name
  • Read and write fields and properties of live objects
dotnet add package com.IvanMurzak.ReflectorNet

Build your own

Start a new extension from the template.

Unity extension template

A ready Unity package layout for your own AI tools and RQL modules.

  • Package layout and an init script
  • Editor and runtime tool examples
  • CI for several Unity versions
git clone https://github.com/IvanMurzak/Unity-MCP.git