Products

MCP Servers & Connectors

MCP servers connect AI models like Claude to the tools and data you already use, over the open Model Context Protocol. These 7 connectors are production-hardened on live client accounts.

Connectors like these exist off the shelf; these are production-hardened from running a live agency: more tools than the basic versions, real-world bug fixes baked in, and capabilities the stock connectors lack. Each one wraps an official API we use every day on client accounts, so the known pitfalls are already fixed inside. They all follow one safety rule: reading data runs freely, but anything that would change an account stops and waits for your approval first. Your API keys stay in your own environment; nothing passes through our servers.

An MCP connector is a small server that gives Claude a live, permissioned link to one external platform, so the model can read and act on that platform's data through the open Model Context Protocol. Each connector on this shelf wraps a single tool, Google Ads, Search Console, Analytics, Meta Ads, ManyChat, Google Business Profile, or Wix, and exposes its real operations to Claude instead of a flat data export. Because we run these on live client accounts, they carry more tools than the stock versions and the bug fixes those accounts forced on us.

MCP stands for Model Context Protocol, the open standard Anthropic released in November 2024 for connecting AI models to outside tools and data. OpenAI adopted it in March 2025 and Google DeepMind followed in April 2025, so the connectors here run on Claude and any other MCP-compatible AI on the same open standard.

7 in the catalog
Who this shelf is for
  • Marketer whose reports still start with screenshots and copy-paste
  • Agency that wants audits running on live account data instead of manual exports
  • Operator connecting Claude to the tools the business already runs on
Where to start

Start with the connector for the platform you check most, for most teams that is Google Search Console or Google Analytics. Connect one, run a single live report, and the screenshot habit ends on its own.

New to this shelf? Read the guides first: What Is MCP? Model Context Protocol Explained

Filter by category

All MCP Connectors

Search & AEO MCP

GBP MCP

Google Business Profile management

$159
Inspect →

Connector categories

The connectors group into four jobs an agency runs every day:

  • Advertising: the Google Ads and Meta Ads connectors let Claude pull spend, campaigns, and performance, then draft changes against the live account.
  • Analytics and search: the Google Analytics and Search Console connectors feed Claude the traffic, query, and conversion data behind every report.
  • Customer messaging: the ManyChat connector opens subscriber flows and conversation data to Claude for messaging automation.
  • Business presence and site: the Google Business Profile and Wix connectors cover the map listing and the site itself, so Claude can read and update where customers actually find you.

Pick the connectors that match the platforms you already run. Each one installs the same way and speaks the same protocol, so adding a second is no harder than the first.

Works with

Universal by design: these run in any AI. Delivered in the open Agent Skills + MCP format (native in Claude); ChatGPT, Gemini, Cursor and Copilot adapt the same files their own way.

  • Claude Native format
  • ChatGPT Adapts via open standards
  • Gemini Adapts via open standards
  • Cursor Adapts via open standards
  • Copilot Adapts via open standards
Questions · still in the air

Catch what's on your mind.

the air is clear. nothing between you and the forge.
catch a spark: the forge will answer

  1. What does an MCP connector actually connect?

    Claude to the real tool: Google Search Console, Analytics, Ads and more, over the official APIs. Reports and audits stop relying on pasted screenshots and start running on live account data.

  2. Is it safe to hand Claude API access?

    The keys stay in your environment, nothing routes through us. The connectors are read-first by design: they pull data freely, and anything that changes an account stays behind your explicit approval.

  3. Do I need to be a developer to set one up?

    No: setup is a config entry and an API key, and each connector ships with the step-by-step. If you can create an API key in the tool’s console, you can wire the connector.

  4. Do these connectors only work with Claude?

    No. MCP is an open standard: Anthropic released it, then OpenAI and Google DeepMind adopted it. The connectors run on Claude and any other MCP-compatible AI on the same protocol.

  5. How are these different from the official connectors?

    They are production-hardened from running a live agency: more tools than the basic versions, real-world bug fixes baked in, and capabilities the stock connectors lack, all wrapping the same official APIs.

  6. How is an MCP connector different from a data-integration platform?

    A data-integration platform moves rows on a schedule: it copies data from one app into a spreadsheet or warehouse and stops there. An MCP connector does something different. It gives an AI model a live, two-way link to the tool itself, so Claude can read the data and also act, draft a campaign change, answer a question, or write a report, in the same session. You are not syncing tables for a human to open later; you are handing the model a permissioned door into the platform. That is why these connectors ship with real operations, not just data exports.