GTM MCP Server

An AI assistant can control your Google Tag Manager containers.

The server connects an AI assistant to the Google Tag Manager API. You give instructions in usual language. The assistant creates tags, examines configurations, writes tracking plans, and publishes changes.

URL: https://mcp.gtmeditor.com

Table of Contents

• Supported AI Clients
• What You Can Do
• Quick Start
  • Claude (Web and Desktop)
  • ChatGPT
  • Gemini CLI
  • Cursor
• Functions
  • Tag Control
  • Trigger Control
  • Container Operations
  • Server-Side Containers
  • Community Template Gallery
  • AI Workflows
• Examples of Use
• How the Server Operates
• Safety Functions
• Comparison with Other GTM MCP Servers
• Self-Hosting
  • Service Account Mode (S2S)
  • Docker Setup
  • Google Cloud Setup
• Available Tools
• Resources and Prompts
• More Context for AI Assistants
• Architecture
• Known Problems
• Links
• Author
• License

Supported AI Clients

The server does not require a specified client. Each MCP client that has OAuth 2.1 with PKCE on an HTTP transport can connect to the server. This includes clients with Dynamic Client Registration (RFC 7591) and clients without it.

What You Can Do

Give these instructions to your AI assistant:

• "List all my GTM containers"
• "Create a GA4 event tag for form submissions"
• "Examine this container for problems and duplicates"
• "Write a tracking plan document for the marketing team"
• "Set up ecommerce tracking for purchases"
• "Publish the changes"

You do not use the GTM interface for these tasks. You do not copy configurations manually. You give the instruction in usual language.

Quick Start

Claude (Web and Desktop)

For Claude.ai:

1. Select Settings, then Connectors, then Add Custom Connector.
2. Type this URL: https://mcp.gtmeditor.com
3. Click Add.
4. Sign in with your Google account.

For Claude Code (CLI), use this command:

claude mcp add -t http gtm https://mcp.gtmeditor.com

ChatGPT

1. Go to the OpenAI Apps Platform.
2. Add an MCP integration with this URL: https://mcp.gtmeditor.com
3. Give permission with your Google account.

Gemini CLI

gemini mcp add --transport http --url https://mcp.gtmeditor.com gtm

Cursor

1. Open Settings, then MCP.
2. Click Add new MCP server.
3. Set the type to URL.
4. Type this URL: https://mcp.gtmeditor.com/authorize
5. Give permission with your Google account.

As an alternative, add this configuration to your .cursor/mcp.json file:

{
  "mcpServers": {
    "gtm": {
      "url": "https://mcp.gtmeditor.com/authorize"
    }
  }
}

Functions

Tag Control

The server creates and changes all GTM tag types:

• GA4 configuration and events. Set up Google Analytics 4 with the correct measurement IDs.
• Ecommerce tracking. Use purchase, add-to-cart, and view-item events.
• Custom HTML. Add scripts, pixels, and custom code.
• Custom image. Add tracking pixels with cache prevention.

Trigger Control

The server creates triggers for these conditions:

• Page views on all pages or on specified URLs
• Custom dataLayer events
• Clicks
• Form submissions
• Timers
• Trigger groups for complex conditions

Container Operations

• Read accounts, containers, and workspaces.
• Create a version from the changes in a workspace.
• Publish a version to make it live.
• Put items in folders.
• Enable or disable the built-in variables.

Server-Side Containers

The server has full support for server-side GTM containers:

• Clients. Create, change, and delete server-side clients. The GA4 client is an example.
• Transformations. Control event parameters with allow, exclude, and augment rules.

Community Template Gallery

The server imports templates from the Google Community Template Gallery. Give instructions such as these:

• "Import the iubenda cookie consent template"
• "Add Cookiebot to my container"
• "Set up Facebook Pixel with the gallery template"

The AI assistant finds the template and its GitHub repository. Then the assistant imports the template automatically.

AI Workflows

Container examination. Give the instruction "Examine my container for problems". The assistant examines the workspace for these conditions:

• Names that do not agree
• Tags that occur more than once
• Triggers that no tag uses
• Risks to security
• Configurations that do not obey the best practices

Tracking plan. Give the instruction "Write a tracking plan". The assistant writes a markdown document with this content:

• All events and their triggers
• The necessary dataLayer values
• The definitions of the variables
• Notes about the implementation

GA4 recommendations. Give the instruction "Help me set up GA4 for ecommerce". The assistant recommends the tags, the triggers, the necessary variables, and the dataLayer code.

Examples of Use

Make a Complete Tracking Setup

The assistant can make a full GA4 ecommerce implementation. Give the instruction "Set up GA4 ecommerce tracking for my store". The assistant then does these steps:

1. It creates 12 tags or more. These include the configuration tag and all the ecommerce event tags.
2. It creates one trigger for each dataLayer event.
3. It creates dataLayer variables for the items, the currency, the value, and the transaction ID.
4. It obeys the Google recommendations for event names and parameters.

Add Consent Control

You can connect privacy tools such as OneTrust to your tracking. Give the instruction "Make GA4 fire only when the user gives consent for analytics". The assistant then does these steps:

1. It creates the variables that read the consent.
2. It creates the related triggers.
3. It changes the applicable tags.

Change Many Items

You can control containers that have many items. Give instructions such as these:

• "Add the prefix 'ecom -' to all ecommerce triggers"
• "Change all tags to use a measurement ID variable"

The assistant can change or move many items in one operation.

Make Custom Variables

You can make complex tracking logic. Give instructions such as these:

• "Create a variable that gives the local time"
• "Add a custom parameter to the purchase tag"

The assistant makes custom JavaScript variables and dataLayer variables.

Use by an Agency

• Control the containers of more than one client.
• Use the same implementation for all clients.
• Set up a new project quickly.
• Make a version and publish the changes safely.

How the Server Operates

The server connects AI assistants to the Google Tag Manager API. It uses the Model Context Protocol. When you give an instruction, the server does these steps:

1. It authenticates you with your Google account through OAuth 2.1.
2. It reads the configuration of your container.
3. It makes the changes that you request.
4. It asks for your approval before a destructive operation.

The server does not keep your Google password. The server uses tokens. You can cancel a token at any time from your Google account.

Safety Functions

• The server asks for approval before a deletion or a publication.
• The server changes only the workspace. No change is live until you publish it.
• The server makes a version before each publication.
• The server writes a log of the changes.

Comparison with Other GTM MCP Servers

More than one MCP server for Google Tag Manager is available. The most usual alternative is [stape-io/google-tag-manage