Google Ads MCP Server

Overview

The Google Ads MCP Server is Google's official open-source implementation of the Model Context Protocol (MCP). It serves as a secure, standardized bridge between Large Language Models (LLMs) and the Google Ads API.

This server allows AI agents to retrieve campaign data, performance metrics, budgets, keywords, and account information using natural language prompts — without writing custom API code or managing complex authentication flows.

Currently read-only (no write operations like creating or editing campaigns), it is actively maintained and optimized for tools like Gemini CLI/Code Assist, with full compatibility for Claude Desktop, Cursor, and other MCP clients.

Key Features

• Natural Language Access: Ask questions like "How is my campaign performance this week?" or "List my top-performing keywords" and get structured answers.
• GAQL Support: Full access to Google Ads Query Language for powerful, custom reporting.
• Multi-Account Support: Tools for listing accessible customers and manager accounts.
• Local & Secure: Runs entirely on your machine with proper OAuth/service account credentials.
• Zero Glue Code: AI clients automatically discover and use tools via MCP standard.
• Python-Based: Easy installation via pipx directly from the GitHub repo.

How It Works

The server runs as a local MCP endpoint (stdio transport). When connected to an AI client:

1. The LLM discovers available tools (search and list_accessible_customers).
2. It generates GAQL queries based on user prompts.
3. The server authenticates via Google Ads API and returns clean, structured results.
4. Results are injected back into the LLM context for analysis and summarization.

Setup & Installation

Prerequisites

• Google Ads Developer Token
• Google Cloud Project with Google Ads API enabled
• OAuth credentials (credentials.json) or service account
• Python + pipx installed

Quick Start

Run via pipx:

pipx run git+https://github.com/googleads/google-ads-mcp.git google-ads-mcp

Configure your MCP client (e.g., Gemini settings.json or Claude/Cursor config) with the required environment variables:

{
  "mcpServers": {
    "google-ads-mcp": {
      "command": "pipx",
      "args": ["run", "--spec", "git+https://github.com/googleads/google-ads-mcp.git", "google-ads-mcp"],
      "env": {
        "GOOGLE_APPLICATION_CREDENTIALS": "/path/to/credentials.json",
        "GOOGLE_PROJECT_ID": "your-project-id",
        "GOOGLE_ADS_DEVELOPER_TOKEN": "your-developer-token"
      }
    }
  }
}

Available Tools

• search: Execute any GAQL query to fetch campaigns, ads, keywords, metrics, budgets, etc.
• list_accessible_customers: Return all accessible customer IDs and names (useful for manager accounts).

Use Cases

• Campaign Analysis: Get instant performance summaries, identify underperforming ads, or compare metrics across accounts.
• Reporting Automation: Ask AI to generate weekly reports or spot trends without logging into the Google Ads dashboard.
• Keyword & Competitor Insights: Natural language queries for keyword performance and optimization ideas.
• Agency Workflows: Let AI agents assist with client reporting and diagnostics across multiple accounts.
• Multi-MCP Orchestration: Combine with other MCP servers (Notion, calendars, analytics) for full-funnel marketing agents.

Limitations & Security Notes

• Read-only mode in the current release (no campaign creation or edits).
• Data is exposed to the connected AI agent/LLM — only connect trusted clients.
• Requires valid Google Ads credentials and developer token.
• Best suited for analysis and diagnostics; full write capabilities may be added in future releases.

Getting Started

Visit the official repository at github.com/googleads/google-ads-mcp for the latest code, issues, and updates. Also check the Google Ads API MCP Developer Guide.

The Google Ads MCP Server is a major step toward agentic advertising — turning your Google Ads data into a conversational, AI-native resource.