Unity-MCP
Unity-MCP is an AI-powered bridge that connects LLMs to the Unity Editor and runtime via the Model Context Protocol, exposing 50+ tools for asset management, scene hierarchy manipulation, C# scripting, and test execution so AI agents can directly develop and debug Unity projects.
Score Breakdown
⚙ Agent Friendliness
🔒 Security
Local Unity Editor integration — no external auth needed for local use. Unity Services (Cloud Build, Analytics) require Unity account credentials separately.
⚡ Reliability
Best When
You are a Unity developer who wants to use Claude Code, Cursor, or another MCP client to accelerate editor workflows, automate asset management, or have an AI pair programmer with direct engine access.
Avoid When
Your Unity project path contains spaces (a hard requirement), or you need a cloud-hosted solution rather than a local MCP server.
Use Cases
- • Using an AI assistant (Claude, Copilot, etc.) to create and manage GameObjects, components, and scenes in Unity
- • Dynamically compiling and executing C# code in the Unity Editor via Roslyn for rapid AI-driven iteration
- • Running Unity EditMode and PlayMode tests from an AI agent and receiving structured results
Not For
- • Non-Unity game engines (Unreal, Godot, etc.)
- • Runtime production builds without careful scoping — reflection access to private methods is a security concern
- • Teams who need a managed cloud service rather than a local MCP server
Interface
Authentication
No authentication required. Local MCP server accessed by MCP client of choice.
Pricing
Apache 2.0 open source. LLM provider costs are separate.
Agent Metadata
Known Gotchas
- ⚠ Project path cannot contain spaces — hard requirement
- ⚠ Reflection access to private methods and compiled DLLs raises security considerations
- ⚠ Supports both stdio and HTTP transports — configuration must match client expectations
- ⚠ Docker support available but adds networking complexity for editor interaction
Alternatives
Full Evaluation Report
Detailed scoring breakdown, competitive positioning, security analysis, and improvement recommendations for Unity-MCP.
Scores are editorial opinions as of 2026-03-06.