Unity

Unity-MCP provides a bridge that lets AI assistants run C# code, inspect Unity scenes and objects, and interact with runtime state through the Model Context Protocol. It enables AI-assisted development, automated testing, scene analysis, and runtime debugging by tying an MCP client to the Unity runtime and persisting results with AILogger.

How to use

You connect your MCP client to Unity-MCP using a JSON configuration that defines how to reach the server. Once connected, you can execute C# code inside the Unity runtime, inspect objects and their components, query scene hierarchies, run tests, and invoke methods on game objects. Results and logs are stored for later retrieval through the integrated AILogger, so you can review what happened during a session.

How to install

Prerequisites: Node.js installed on your development machine. AILogger should be running at http://localhost:3030 or you should set the AI_LOGGER_URL environment variable to point to your AILogger instance.

1. Clone the project repository

2. Change into the project directory

3. Install dependencies

4. Build the project

5. Start the MCP STDIO client which talks to Unity and uses AILogger for persistence

# 1. Clone the repository
git clone https://github.com/TSavo/Unity-MCP.git
cd Unity-MCP

# 2. Install dependencies
npm install

# 3. Build the project
npm run build

# 4. Start the MCP STDIO client
npm start

Configuration and notes

Unity-MCP exposes both an HTTP-based endpoint and a local STDIO workflow for connecting with MCP clients. The HTTP endpoint starts a server you can reach from your MCP client, while the STDIO workflow runs the MCP STDIO client locally and communicates with Unity through the Unity client. Ensure AILogger is available at the configured URL to persist logs and results.

Available tools

execute_code

Execute C# code directly in the Unity runtime environment and store results in AILogger for retrieval.

query

Query Unity objects and properties using dot notation to access the runtime state.

get_logs

Retrieve logs stored by AILogger for auditing or debugging.

get_log_by_name

Retrieve a specific log entry from AILogger by its log name.