Google's official routes into Google Ads data: the Google Ads API, and an open source MCP server that lets an AI assistant read account data.
- Kind
- Connector
- Keeplight status
- Researched, not yet tested
- Cost
- As of 2026-11: Google's guide lists no fee for the server. Ad spend is billed as usual.
- Connects to
- Google Ads, Google Cloud, Google Ads API
- Official site
- developers.google.com/google-ads/api/docs/developer-toolkit/mcp-server
What it does
The Google Ads API is the official way for software to read and manage Google Ads accounts. Google also publishes a Google Ads MCP server, which lets an AI assistant work with the API in plain language.
Google's guide describes the MCP server as read and partial change. It can list accounts, run queries on account data, create assets, upload videos in local mode, and manage labels. It cannot create or change campaigns, ad groups, ads, budgets, or bids. Changes use a two-step propose and confirm process.
It suits a team that wants reporting and checks on Google Ads without logging in to the interface. Setup uses Google Cloud and OAuth, so expect an hour or two for the first connection.
Use cases
Ask which campaigns spent the most
Connect the MCP server to your AI assistant and ask for spend by campaign. It runs a query against the account and returns the rows.
Find the right field before writing a report
The metadata tool looks up which fields can be queried, which saves guessing when you build a report.
Keep labels tidy
The label tools add or remove labels on ads and campaigns, so a team can mark what is under review. Changes ask for confirmation first.
Feed a weekly report with a scripted query
A script can call the Google Ads API on a schedule and save the results to a sheet that a report reads.
How to start
- Open Google Cloud, create a project, and enable the Google Ads API.
- Check which API access level your account has. Production use needs Explorer, Basic, or Standard access.
- Install Python 3.10 or newer, pipx, and the gcloud command-line tool.
- Run gcloud auth application-default login with the Google Ads API scope and your OAuth client file.
- Add the google-ads-mcp entry from Google's guide to your AI tool's MCP settings and set your project ID.
- Start with a read-only question about one account, and turn off the labels and assets tools if you want it strictly read-only.
Visit Google Ads API and MCP server
Researched, not yet tested: Keeplight has read the docs but has not run it yet. Prices and features change, so check the official site before you buy.