Blog
A Model Context Protocol server lets AI agents create tags, triggers, and variables in your container conversationally.
The SprTags MCP server exposes container management as a set of tools that Claude (or any MCP-compatible AI agent) can call directly. Instead of clicking through the GTM UI for routine tasks, you describe what you want and the agent does it.
Model Context Protocol is an open standard for connecting AI agents to external tools. An MCP server publishes a set of tools (name, description, parameters); an MCP-aware client connects, discovers the tools, and calls them when the user's request matches.
SprTags ships an MCP server with tools for listing accounts, creating workspaces, and managing tags, triggers, variables, and versions in any GTM container linked to your dashboard.
In your SprTags dashboard, link a Google account that has GTM access. Enable the MCP server. Generate an MCP token. Add the token to your AI client's MCP configuration (Claude Code, Claude Desktop, Cursor, or any other MCP-aware tool).
Any tool that deletes or publishes (delete container, delete tag, publish version) is blocked unless the agent passes confirm: true. This is intentional. Treat it as a safety check; always verify the change before confirming.
Audit prep: ask the agent to list tags last edited more than six months ago. The output goes straight into your audit pass.
Bulk variable creation: when a new event taxonomy adds 12 parameters, asking the agent to create the matching variables saves 30 minutes of clicking. Each variable gets the right name, type, and data layer reference.
Cross-container comparisons: ask the agent to compare two containers and list the tags present in one but not the other. Useful for staging-to-production drift checks.
For one-off changes you would do manually in 30 seconds, the MCP setup overhead is not worth it. For repetitive operations or for asking analytical questions about your container's state, it is much faster than the UI.