Free AI Game Developer | Unity MCP

# AI Game Developer > AI Game Developer is an MCP (Model Context Protocol) platform that lets AI agents inspect and drive real game engines. You install one editor plugin — Unity-MCP, Unreal-MCP, or Godot-MCP — into your project, and it exposes that engine's editor operations as AI Tools. Any MCP-compatible AI client (Claude, Cursor, GitHub Copilot, Gemini, Codex, Windsurf, Cline) can then create scenes, spawn objects, edit and compile scripts, manage assets, and capture screenshots directly inside your live editor. The plugin is the bridge, not an AI model: you bring the AI agent, AI Game Developer gives it hands inside the engine. It is free and open source, and connects through the hosted ai-game.dev cloud backend or a self-hosted GameDev-MCP-Server. ## Start here - [Home](https://ai-game.dev/): What AI Game Developer is and which engines it supports. - [Download](https://ai-game.dev/download): Get the plugin / installer. - [Quickstart](https://ai-game.dev/docs/quickstart): Install, authorize, and run your first AI request in minutes. - [MCP endpoint info](https://ai-game.dev/mcp-endpoint-info): Per-client MCP config (Cloud vs Custom / self-hosted). ## Engines - [Unity](https://ai-game.dev/engines/unity): Unity-MCP — 190 tools across 27 families (76 core + 114 from extensions); works in Editor and Runtime. Unity 2022.3+. - [Unreal Engine](https://ai-game.dev/engines/unreal): Unreal-MCP — 62 tools across 8 families incl. Blueprint and C++ authoring/compile. UE 5.5+ (tested on 5.7). - [Godot](https://ai-game.dev/engines/godot): Godot-MCP — 36 tools across 10 families, C# and GDScript. Godot 4.3–4.5 (mono). ## Best practices How to make simple, effective AI requests and discover what the agent can do — per engine. - [Best practices hub](https://ai-game.dev/docs/best-practices): Principles, the anatomy of a good request, and capability-discovery recipes (engine-agnostic). - [Unity best practices](https://ai-game.dev/docs/best-practices/unity): E