AI ads infrastructure

What Is MCP for Google Ads?

MCP for Google Ads is the open-standard interface that lets AI assistants like Claude read and act on your campaigns. Plain-English explainer + setup.

NotFair Team|

MCP for Google Ads is an MCP (Model Context Protocol) server that exposes Google Ads API operations as structured tools an AI assistant can call. Plug it into Claude, ChatGPT, Cursor, or any MCP-compatible client and the assistant can read campaigns, run audits, pause keywords, adjust bids — all from a normal conversation.

The Model Context Protocol itself is an open standard from Anthropic that defines how AI clients connect to tools. Think of it as USB-C for AI. "MCP for Google Ads" is a specific server implementation that speaks the protocol on one side and the Google Ads API on the other.

What does an MCP for Google Ads do?

  • Translates natural language into Google Ads API calls"audit my account" becomes a parallel GAQL fan-out across spend, search terms, and quality scores.
  • Holds the Google Ads developer token (hosted servers) so individual users do not need to apply to Google.
  • Manages OAuthone click to grant read/write scope; revocable through Google's account settings.
  • Enforces approval gates on destructive operationspausing campaigns, deleting assets, big budget changes.
  • Exposes both read tools (audits, reports, change history) and write tools (pause, bid update, negative keywords, RSAs).

Why MCP and not just the Google Ads API directly?

The Google Ads API is built for human developers writing code. MCP is built for AI clients picking tools at runtime. The MCP layer makes Google Ads accessible to any compatible AI without per-client integration code — connect once, use anywhere.

MCP also adds the safety layer the Google Ads API does not have on its own: approval flows, scoped auth, change history, and rollback. The raw API will let an AI delete a campaign with a single call. A well-designed MCP server requires confirmation first.

How do you actually use one?

Three steps: pick an MCP server (the simplest is NotFair, hosted), authenticate with Google through the server's OAuth flow, and add the server to your AI client. From that point on, you can ask "why did CPA rise this week?" or "pause keywords with zero conversions in 90 days" and the AI will route the call to the right tool.

FAQ

Try MCP with Google Ads

Connect your Google Ads account to NotFair in 30 seconds and start querying campaigns from Claude.

Connect Google Ads

FAQ

Common questions about Model Context Protocol.

Is MCP for Google Ads the same as Google's AI features?

No. Smart Bidding, Performance Max, and other Google-built AI features run inside the Google Ads platform. MCP for Google Ads is what lets external AI assistants — Claude, ChatGPT, Cursor — read and act on your account. They are complementary, not competing.

Does MCP work with Microsoft Ads or Meta Ads?

Yes — separate servers exist or are being built for those platforms. MCP is a generic protocol; each ad platform needs its own server implementation.

Is MCP for Google Ads free?

The MCP protocol is free and open. NotFair's hosted server is free during the public beta. Community open-source servers are free but require you to bring your own Google developer token.

What client should I use to connect MCP for Google Ads?

Claude Desktop is the easiest entry point — one-click install via the Connectors UI. Claude Code, ChatGPT (via Codex), and Cursor all also work.