google_ads_mcp_server

This repo has been archived in favor of: https://github.com/googleads/google-ads-mcp

Installation

Run this command in your terminal to add the MCP server to Claude Code.

Run in terminal:

Command

claude mcp add --transport stdio google-marketing-solutions-google_ads_mcp_server uvx --directory path/to/google_ads_mcp run server \
  --env USE_PROTO_PLUS="true" \
  --env JSON_KEY_FILE_PATH="Path to service account JSON key" \
  --env GOOGLE_ADS_DEVELOPER_TOKEN="Your developer token" \
  --env GOOGLE_ADS_LOGIN_CUSTOMER_ID="Your login customer ID"

How to use

This Google Ads MCP server provides a standardized way to integrate a large language model (LLM) with Google Ads by exposing a small set of tools that query and inspect Google Ads data. The server offers at least two tools: list_accounts, which lists all available Google Ads accounts, and search_stream, which runs a GAQL query against a specified customer_id. You can use these tools to fetch account metadata, discover available campaigns or queries, and feed results into your LLM-driven workflows. To use the server with Gemini or other MCP tooling, point your MCP client to the googleAdsMcpServer configuration and invoke the run server command to start the local or remote MCP service, allowing your agent to call the available tools and receive structured results.

How to install

Prerequisites:

Python 3.8+ installed on your system
uv (Python package) installed or accessible via your environment
Access to Google Ads API with a service account and the appropriate credentials

Installation steps:

Install the required Python/MCP tooling (example using uvx):

pip install uvx
Prepare Google Ads credentials:
- Create a Google Cloud project and enable Google Ads API
- Create a service account and download the JSON key
- Place the key and configuration in a secure location
- Create a google-ads.yaml configuration file as described in the Google Ads API docs, and ensure the file contains the necessary fields (developer_token, login_customer_id, json_key_file_path, use_proto_plus)
Clone or download the google_ads_mcp_server repository to a directory on your machine.
Start the MCP server using uvx (or the equivalent command in your environment):

uvx --directory path/to/google_ads_mcp run server
Verify the server is running and reachable via your MCP client or Gemini CLI configuration.

Note: Replace path/to/google_ads_mcp with the actual path to your downloaded/google_ads_mcp_server codebase.

Additional notes

Environment variables and configuration:

Ensure GOOGLE_ADS_DEVELOPER_TOKEN, GOOGLE_ADS_LOGIN_CUSTOMER_ID, and JSON_KEY_FILE_PATH are correctly set in your environment or in the google-ads.yaml file.
The USE_PROTO_PLUS flag can be toggled depending on whether your client expects proto_plus support.
If you run into authentication or quota errors, verify that the service account has the necessary Google Ads API permissions and that the developer token is valid.
When integrating with Gemini CLI or other MCP clients, ensure the MCP server's directory path is correctly provided in the client settings (as shown in the example in the README).