GOOGLE ADS MCP SERVER · READ-ONLY · OAUTH

Google Ads MCP server for your AI assistant.

Your Google Ads data. A better conversation. Bring campaign evidence into your AI assistant. Explore search intent, understand changes in spend and check conversion settings before deciding what to do next.

The hosted Google Ads MCP server needs no API keys, no local process and no developer setup: paste one endpoint into your AI client, sign in with Google, and ask.

Connect your assistant

What you can do

Find the right account

Discover accessible customers and expand manager accounts. Read currency and timezone before comparing results.

Understand performance

Read campaign, keyword, search-term, landing-page and device reports over explicit date ranges.

Check the configuration

Inspect budgets, ads, conversion actions and recommendations. Use bounded GAQL queries for deeper analysis.

SETUP GUIDE

From connection to first answer.

You need an MCP client with remote HTTP and OAuth support, plus a Google login with access to the data you want to read. Client availability can depend on your plan or administrator.

Add a remote MCP server
Open your AI client’s MCP or connector settings. Add a server named “Adlane Google Ads” and paste the URL below. Choose Streamable HTTP if your client asks for a transport.
https://mcp.adlane.app/gads
Sign in and review consent
Start the connection in your client and follow the browser sign-in. Review the requested access, approve if appropriate, then return to your AI client.
Google uses its broad adwords permission for API access. Adlane exposes only read-only tools through this MCP. The shared Google consent screen may display Liftaven.
Check the tools
Refresh the connection if needed. You should see 13 tools, including list_accessible_customers. Ask your assistant to discover accounts before requesting a report.
Ask your first question
Start with a specific scope. Copy or adapt this prompt:
List my accessible Google Ads accounts. Ask me which customer to analyse, then check its currency, timezone and conversion actions. Compare the last 28 complete days with the previous 28 days. Explain changes in spend and conversions, cite the reports and flag missing data. Do not change anything.

Set up the Google Ads MCP server in Claude, Claude Code, Cursor or VS Code

Every client below supports remote HTTP MCP with OAuth. The server name is yours to choose; adlane-google-ads is used in the examples.

Claude (web and desktop)

Open Settings, choose Connectors, then Add custom connector. Paste https://mcp.adlane.app/gads as the remote MCP server URL and save. Click Connect on the new connector to start Google sign-in.

Claude Code

Run the command below in your terminal, then type /mcp inside Claude Code to authenticate.

claude mcp add --transport http adlane-google-ads https://mcp.adlane.app/gads

Cursor

Add this entry to .cursor/mcp.json in your project or to the global MCP settings. Cursor opens the browser for OAuth the first time the server is used.

{
  "mcpServers": {
    "adlane-google-ads": {
      "url": "https://mcp.adlane.app/gads"
    }
  }
}

VS Code (GitHub Copilot agent mode)

Create .vscode/mcp.json with this entry, then start the server from the MCP view.

{
  "servers": {
    "adlane-google-ads": {
      "type": "http",
      "url": "https://mcp.adlane.app/gads"
    }
  }
}

Desktop clients without remote OAuth support

If your client accepts an mcpServers JSON configuration but cannot connect to a remote OAuth server directly, the third-party mcp-remote bridge can make the connection. Install Node.js with npm first, merge this entry into your configuration, restart the client and complete browser sign-in. Native remote MCP support is preferable when your client provides it. Do not paste provider passwords or access tokens into any configuration.

{
  "mcpServers": {
    "adlane-google-ads": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.adlane.app/gads"
      ]
    }
  }
}

Available Google Ads MCP tools

These tools read data. They cannot create campaigns, change budgets, publish ads or approve Adlane proposals.

list_accessible_customersget_customerlist_customer_clientslist_campaignslist_budgetslist_ad_groupslist_adslist_conversion_actionslist_recommendationslist_asset_groupsrun_reportsearchsearch_fields

Search-term reports omit privacy-filtered queries. Conversion value is not profit, and attributed conversions can change after the report date. Access also depends on Google Ads API approval and your account permissions.

Hosted Google Ads MCP server vs running your own

Open-source Google Ads MCP servers on GitHub suit developers who are comfortable creating API credentials and running a local process. The hosted endpoint trades that control for a two-minute setup, read-only scope and nothing to maintain.

Consideration	Self-hosted open-source server	Adlane hosted endpoint
Setup	Register an app or developer token, create OAuth credentials, install and run locally	Paste one URL into your client and sign in with Google
Credentials	Stored on your machine; you rotate and secure them	Tokens encrypted server-side and bound to your connection; revoke from your Google account at any time
Works from	The computer running the process	Any client on any device, including web-based clients
Write access	Depends on the project; some expose campaign and budget mutations	None. Changes exist only as reviewed proposals inside the Adlane workspace
Cost	Free software; your time to set up and maintain	Free, with no Adlane subscription required

Both approaches read the same provider data. Choose self-hosting if you need custom tools or write access under your own controls; choose the hosted endpoint if you want Google Ads evidence in your AI client today without touching an API console.

Google’s official Google Ads MCP server vs Adlane

Google’s Ads API team publishes an official, open-source Google Ads MCP server. It is a good choice if you already have a Google Cloud project with Google Ads API access and want to run and audit the code yourself. Like Adlane’s endpoint, the current release only reads data.

The difference is who operates it and how much of the API it wraps. The official server gives your assistant raw GAQL search; Adlane adds report tools that convert cost micros, read currency and timezone first, and expand manager accounts.

Consideration	Google’s official server	Adlane hosted endpoint
Who runs it	You, locally (stdio) or on your own Cloud Run service	Adlane, at a fixed HTTPS URL
What you need	Python, a Google Cloud project with Google Ads API access, an OAuth client or service account	A Google login that can already see the account
Access	Read-only in the current release	Read-only
Tools	list_accessible_customers, search (GAQL), get_resource_metadata	13 tools, including campaign, search-term, budget and conversion-action reports
Money values	Raw micros, as the API returns them	Report tools return account-currency units

Source: Google Ads API: MCP server. Check the provider’s current documentation before you rely on a feature list; these products change quickly.

Connection troubleshooting

No accounts or data?

Sign-in did not finish?

Complete consent in the same browser session. If a desktop bridge uses a local return address, allow that callback in your browser. Reconnect from the client if the sign-in expires.

Access stopped working?

Reconnect after access is revoked or a provider token expires. For Meta, expired tokens require a new sign-in. Check your provider permissions before repeatedly retrying a failed report.

Google Ads MCP server: questions, answered

Is this the official Google Ads MCP server?

No. Google publishes an open-source, read-only Google Ads MCP server that you run yourself, locally or on Cloud Run, with a Google Cloud project that has Google Ads API access and your own OAuth client or service account. Adlane’s is a hosted, read-only endpoint: paste one URL, sign in with Google, and there is nothing to install or rotate.

Do I need a Google Ads developer token?

No. Adlane holds the API application credentials server-side. You authorise access to the customer accounts your Google login can already see.

Can it create campaigns or change budgets?

No. The MCP tools read accounts, campaigns, budgets, conversion actions and reports. Changes happen only inside the Adlane workspace, as proposals a human reviews and approves.

Why do I see a manager account but no campaigns?

A manager account holds client accounts; campaign data lives in the client. Ask your assistant to list customer clients, then read reports for the right customer ID.

Why are cost figures a million times too large?

The Google Ads API reports money in micros. Divide by one million to get account-currency units. Adlane’s report tools do this for you; raw GAQL searches return micros as the API does.

Which clients are supported?

Any client with remote HTTP MCP and OAuth: Claude, Claude Code, Cursor and VS Code are covered in the setup guide. ChatGPT needs a plan that allows custom connectors.

Does it cost anything?

The MCP endpoint has no Adlane subscription gate. Your AI client may have its own plan requirements, and Google Ads API access depends on your account and Google’s approval.

Know what leaves your account.

Adlane keeps provider credentials server-side. Requested report data is returned to your chosen MCP client, which may share it with its AI provider. Review that client’s privacy settings before using sensitive reports.

Read Adlane’s privacy notice