Popular seller tasks

Google Ads MCP

Query advertising account information through the search tool

Open sourceAny store

Overview

Google’s official MCP server exposes Google Ads reporting tools to an assistant. Its documented tools focus on queries, resource metadata and accessible customer accounts.

Who it suits

Ecommerce advertising analysts who need to query a store’s Google Ads results, including eligible Shopping-related reporting, and can configure access to the correct customer account.

Who it may not suit

Users looking for product-feed repair or an autonomous bid-changing system; the documented tool set centers on search, metadata, and accessible-customer discovery.

Seller use cases

  • Query a store’s Shopping advertising results for a defined period to investigate spending without matching conversion value.
  • Read resource metadata before constructing product or campaign reporting queries.
  • List accessible customer accounts and select the correct merchant account for a recurring performance review.

What you can do

  • Query advertising account information through the search tool. ↗
  • Inspect resource metadata before composing report queries. ↗
  • List customer accounts accessible to the authenticated user. ↗

What you need

  • Python tooling, Google Cloud setup and Google Ads authorization for the target account.

Install

Claude Desktop

{
  "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"
      }
    }
  }
}
Installation source ↗

Permissions & data

The Python server uses Google credentials and contacts Google Ads APIs. OAuth deployments can persist encrypted token state in the configured storage.

Project info

Maintainer
googleads
GitHub stars at check date
1096
Last commit
2026-10-09
License
Apache-2.0
Last checked
2026-10-30
GitHub ↗

Pricing at a glance

Google publishes the server under Apache-2.0. Running the Python connection requires Google Ads authorization and any applicable developer access; the software license does not include ad spend or the AI assistant interpreting the reports.

Open source

Good to know

A report query needs compatible metrics and segments. Inspect resource metadata, identify the customer and reporting window, and reconcile conversion definitions before interpreting cost or return figures as store profit.

Alternatives to Google Ads MCP

Claude Marketing provides paid-media audit methods from supplied or connected evidence; this server is the Google Ads data connection itself.

Merchant Center MCP diagnoses product-feed and eligibility issues; Google Ads MCP explains advertising performance around those products.

Pipeboard handles Meta advertising and exposes supported campaign writes, whereas Google’s documented MCP focuses on Google Ads queries.

Frequently asked questions

What makes this useful to ecommerce sellers?

It brings the seller’s Google Ads reporting into an assistant, including relevant Shopping campaign data available through the API. The analysis must use the account’s actual conversion setup.

Can it repair Merchant Center product data?

Merchant Center feed resources are handled by a different API. Use a Merchant Center connector for product issues and this server for advertising reporting.

Does the documented server change bids?

The reviewed tool list contains search, resource metadata, and accessible-customer discovery. Do not assume that the wider Google Ads API’s mutation methods are exposed as MCP tools.

Can I enable only selected reporting tools?

The repository documents tools_config.yaml namespaces and individual-tool controls. Restricting the configured tool set can keep a reporting client focused on the queries and metadata it actually needs.

Checked against the official site on 2026-10-30 · Official source ↗

Sources