GladeKit MCP

Connect Cursor, Claude Code, Windsurf, Claude Desktop, and VS Code to your Unity Editor.

222+ tools. A full Unity-aware system prompt. GLADE.md project context. Script semantic search. Skill calibration. Cloud intelligence layer with RAG and cross-session memory. All core features are free and local — no GladeKit account required.

---

Quick Start

1. Install the Unity package

In Unity: Window > Package Manager > + > Add package from git URL...

https://github.com/Glade-tool/glade-mcp-unity.git?path=/unity-bridge

The Unity bridge starts automatically on localhost:8765.

2. Connect your AI client

Install uv (one-time): curl -LsSf https://astral.sh/uv/install.sh | sh

Then add the MCP config to your AI client. The client launches the MCP server automatically — no manual server step.

Tip: Open Window > GladeKit MCP in Unity to see connection status, or click "Copy MCP Config to Clipboard" for the JSON snippet.

<details> <summary><strong>Claude Code</strong></summary>

If you cloned this repo, the .mcp.json auto-connects. Otherwise add to your Claude Code MCP settings:

{
  "mcpServers": {
    "gladekit-unity": {
      "command": "uvx",
      "args": ["gladekit-mcp"]
    }
  }
}

</details> <details> <summary><strong>Cursor</strong></summary>

Cursor Settings > MCP > Add new MCP server:

{
  "mcpServers": {
    "gladekit-unity": {
      "command": "uvx",
      "args": ["gladekit-mcp"]
    }
  }
}

</details> <details> <summary><strong>Claude Desktop</strong></summary>

Edit ~/Library/Application Support/Claude/claude_desktop_config.json (Mac) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "gladekit-unity": {
      "command": "uvx",
      "args": ["gladekit-mcp"]
    }
  }
}

</details> <details> <summary><strong>Windsurf</strong></summary>

Edit ~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "gladekit-unity": {
      "command": "uvx",
      "args": ["gladekit-mcp"]
    }
  }
}

</details> <details> <summary><strong>VS Code (GitHub Copilot)</strong></summary>

Add to .vscode/mcp.json in your workspace:

{
  "servers": {
    "gladekit-unity": {
      "type": "stdio",
      "command": "uvx",
      "args": ["gladekit-mcp"]
    }
  }
}

</details>

---

Why GladeKit MCP?

Feature	GladeKit MCP	unity-mcp (CoplayDev)
Tools	222+ granular tools across 24 categories	~40 consolidated tools
System prompt	Full Unity intelligence — render pipeline detection, input system routing, tool discipline rules	None
Project context	GLADE.md — inject your game design doc into every request	None
Script search	Semantic search via OpenAI embeddings (bring your own key)	None
Skill calibration	Auto-detects beginner/expert, adapts response verbosity	None
In-session memory	remember_for_session — AI stores and recalls facts mid-conversation	None
Cloud intelligence	GLADEKIT_API_KEY — RAG knowledge base + cross-session memory + conventions	None
License	MIT	MIT

All core features are free and local — no GladeKit account needed. The cloud intelligence layer requires a GLADEKIT_API_KEY from gladekit.com.

---

Features

222+ tools across 24 categories

Scene Management - GameObjects - Hierarchy - Transform - Scripts - Prefabs - Materials - Lighting - VFX & Audio - Animation - Physics - Camera - UI - Input System (New) - Input System (Legacy) - Components - Assets - Import Settings - Selection - Terrain & NavMesh - Project Settings - Utility

All 222+ tools are dispatchable. Claude Code sees 82 curated tools by default (Claude Code has a practical 128-tool limit). Use get_relevant_tools to discover extended tools for specialized work (blend trees, NavMesh, IK, Cinemachine, etc.).

4 meta-tools

• get_relevant_tools — describe what you want to do, get the right tools for the job. Also injects RAG context from the GladeKit knowledge base (paid tier).
• remember_for_session — store a fact for the current conversation. With GLADEKIT_API_KEY, facts are also persisted to the cloud for future sessions.
• recall_session_memories — retrieve all facts stored during this session.
• search_project_scripts — semantic search across your project's scripts (requires OPENAI_API_KEY).

7 resources

The server exposes Unity project state as MCP resources: bridge health, project context, scene hierarchy, project scripts, current selection, GLADE.md, and session memory.

1 prompt

unity-assistant — a full Unity development system prompt that auto-detects your render pipeline (URP/HDRP/Built-in), input system (new/legacy), and injects GLADE.md, skill level, and cloud memories.

GLADE.md — project game design doc

Create a GLADE.md file in your Unity project root. The MCP server reads it and injects it into every request. Works as a permanent context layer: your game's design intent, conventions, and constraints are always in scope.

# My Game

Genre: 2D platformer  
Player: CharacterController, double jump enabled  
Art style: pixel art, 16x16 sprites  
Naming: PascalCase for scripts, snake_case for folders

Script semantic search

With OPENAI_API_KEY set, the server finds relevant scripts using cosine similarity rather than filename matching. Ask "how does the enemy spawn?" and the right script surfaces — even if it's not named EnemySpawner.

Skill calibration

The server tracks vocabulary across your messages and detects whether you're a Unity beginner or expert. Beginners get plain-language explanations and encouraging framing. Experts get terse, technical responses. Calibration persists to .gladekit/skill_level.json in your project.

Cloud intelligence — GLADEKIT_API_KEY

Set GLADEKIT_API_KEY to unlock the paid intelligence layer:

• RAG knowledge base — curated Unity docs, API corrections, and error patterns injected alongside tool recommendations via get_relevant_tools. Powered by pgvector semantic search.
• Cross-session memory — facts stored with remember_for_session are persisted to the cloud and loaded into future sessions. The AI remembers your project context across conversations.
• Convention extraction — coding patterns distilled from your accumulated session memories, surfaced automatically in the system prompt.

Get your API key at gladekit.com. Without the key, all core features still work — the intelligence layer is additive.

---

Environment Variables

Variable	Required	Description
UNITY_BRIDGE_URL	No	Unity bridge URL (default: http://localhost:8765)
OPENAI_API_KEY	No	Enables script semantic search
GLADEKIT_API_KEY	No	Enables RAG knowledge base + cross-session memory + conventions
GLADEKIT_CLOUD_URL	No	Override cloud API URL (default: https://api.gladekit.com/functions/v1)

---

Troubleshooting

Bridge not connecting

• Open Unity and wait for it to finish importing assets — the bridge starts automatically
• Check Window > GladeKit MCP in Unity — the Bridge and AI Client indicators show connection status
• Verify nothing else is using port 8765: lsof -i :8765 (Mac/Linux) or netstat -ano | findstr 8765 (Windows)

AI client can't find uvx

• Install uv: curl -LsSf https://astral.sh/uv/install.sh | sh (Mac/Linux) or pip install uv
• Or use pip install gladekit-mcp and change the config command from "uvx" to "python" with args ["-m", "gladekit_mcp"]

Tools not appearing in Claude Code

• Claude Code has a practical ~128-tool limit. GladeKit shows 82 curated tools by default — this is intentional. All 222+ are dispatchable: use the get_relevant_tools tool to find extended tools by task description.

GLADE.md not being picked up

• The file must be named exactly GLADE.md (case-sensitive on Mac/Linux) and placed in the Unity project root (same directory as Assets/, Packages/, ProjectSettings/)

Cloud features not working

• Verify GLADEKIT_API_KEY is set in your environment (the MCP server logs "GladeKit cloud features enabled" on startup if detected)
• Check that your API key is valid and your plan is active at gladekit.com
• Cloud features require a paid plan. Free tier users see a 403 response.

---

Architecture

[AI Client: Cursor / Claude Code / Windsurf / Claude Desktop / VS Code]
         |
         | stdio MCP protocol
         v
[gladekit_mcp Python process]
    bridge.py    -> HTTP localhost:8765 (Unity bridge)
    prompts.py   -> system prompt (render pipeline, input system, GLADE.md, skill level, memories)
    tools/       -> 222+ tool schemas + dispatch
    cloud.py     -> GLADEKIT_API_KEY -> api.gladekit.com (RAG + memory)
    search.py    -> OPENAI_API_KEY -> script semantic search
    skill.py     -> skill calibration (beginner/expert detection)
         |
         | HTTP localhost:8765
         v
[Unity Bridge — C# Editor extension (UPM package)]
    UnityBridgeServer.cs      -> HttpListener on :8765
    222+ ITool implementations -> auto-discovered via reflection
    UnityContextGatherer.cs   -> scene, scripts, packages, render pipeline
    BackupManager.cs          -> undo/backup before destructive operations

---

Contributing

The Unity bridge (unity-bridge/) is the source of truth for C# tools. Adding a tool requires two files:

1. C# implementation (unity-bridge/Editor/Tools/Implementations/<Category>/MyTool.cs):

public class MyTool : ITool
{
    public string Name => "my_tool";
    public string Execute(Dictionary<string, object> args)
    {
        // ... Unity Editor API calls ...
        return ToolUtils.CreateSuccessResponse("Done", extras);
    }
}

2. Python schema (mcp-server/src/gladekit_mcp/tools/<category>.py):

Add an entry to the category's tool list following the existing format (OpenAI function-calling schema).

Tools are auto-discovered via reflection — no registration needed beyond these two files.

---

License

MIT: see LICENSE.

The GladeKit desktop app provides the full agentic loop experience (streaming, miss recovery, error tracking, memory UI) and is a separate commercial product.