Model Context Protocol (MCP) server overview

The Model Context Protocol (MCP) is a standard that lets AI apps talk to tools like APIs, ad platforms, databases, and analytics platforms in a consistent way. It makes it possible for AI models to understand and perform tasks like “create a new Google Tag Manager container” or “list tags” without needing code or manual API calls.

The MCP server is Stape’s proprietary engine behind this. It connects your AI assistant (like Claude or ChatGPT) to Google Tag Manager (GTM) or server-side containers within Stape. The MCP server takes your plain-language requests and translates them into real actions on these platforms.

There are four ways you can use an MCP server:

• For Stape: a simple way to manage containers on Stape via AI.
• For GTM: manage containers in your account, or tags, triggers, and variables in your GTM containers.
• For GA4: a method provided by Google to allow you to retrieve information from Google Analytics 4.
• For Google Ads: manage and analyze your Google Ads campaigns via AI.

1. 1.
   Manage sGTM containers via chat. For example, you can say "Create a new server container for client-x.com on Stape" instead of clicking through the dashboard.
2. 2.
   Bulk container administration for agencies. An agency managing dozens of client sGTM containers can ask an AI assistant to list, rename, or delete containers across accounts in seconds.
3. 3.
   Onboarding automation. New client onboarding scripts (such as "set up a container, name it X, and confirm it's running") that a non-technical person can trigger.

1. 1.
   Tag deployment. For example, you can say "Create a GTM tag that fires a Facebook Pixel PageView event on all pages" – the AI creates the tag, trigger, and variable automatically.
2. 2.
   Multi-workspace management. You can create, modify, or delete GTM accounts, workspaces, containers, and environments straight from a chat prompt – useful for freelancers who set up many client containers.
3. 3.
   Template and folder organization. You can ask AI to group tags into folders or apply custom templates, keeping large containers tidy without manual drag-and-drop.

1. 1.
   On-demand reporting. For example, you can say "Show me last month's conversion rate by traffic source" returns a live GA4 report directly in the chat, no need to open GA4 dashboard.
2. 2.
   Real-time monitoring. You can pull real-time user activity/event data to catch anomalies (for example, a tracking break or traffic spike) as they happen.
3. 3.
   Metric querying. You can ask custom questions like "compare bounce rate between mobile and desktop for Q2" without writing a GA4 API query.

1. 1.
   Campaign performance analysis. For example, you can say "What's the total spend across all campaigns this month?" and get a structured summary with key metrics – no need to navigate the Google Ads dashboard manually.
2. 2.
   Underperforming campaign detection. You can ask "Which campaigns have the lowest ROAS this week?" to quickly identify where to cut the budget or adjust targeting.
3. 3.
   Cross-period comparison. You can pull and compare campaign results across time periods with a simple prompt like "Compare conversions from last month vs. this month by campaign."