GOOGLE ADS
Google Ads MCP Server Setup Guide — Connect Claude AI to Google Ads API in 2026
The Google Ads MCP Server enables direct Claude AI integration with Google Ads API through Model Context Protocol. Complete setup takes under 15 minutes — from Google Cloud configuration to live GAQL queries in Claude Desktop. This open-source implementation supports OAuth 2.0, automatic token refresh, and 7 automation workflows.
Contents
Autonomous Marketing
Grow your business faster with AI agents
- ✓Automates Google, Meta + 5 more platforms
- ✓Handles your SEO end to end
- ✓Upgrades your website to convert better
What is the Google Ads MCP Server?
The Google Ads MCP Server is an open-source implementation of the Model Context Protocol (MCP) that creates a direct bridge between Large Language Models like Claude AI and the Google Ads API. Released by Google's Marketing Solutions team in 2024, this experimental tool eliminates the need for manual CSV exports, copy-pasting data, or switching between browser tabs when analyzing Google Ads performance.
Model Context Protocol is Anthropic's standard for connecting AI models to external data sources and tools. The google ads mcp server specifically enables Claude to pull live campaign data, execute GAQL (Google Ads Query Language) queries, analyze performance metrics, and recommend optimizations — all through natural language prompts. Instead of spending 3-4 hours per week manually building reports, marketers can ask Claude questions like "Which campaigns have CPA > $50?" and get structured answers in seconds.
The server handles OAuth 2.0 authentication automatically, manages token refresh, and provides secure access to your Google Ads accounts without exposing API credentials. It supports all major Google Ads entities: campaigns, ad groups, ads, keywords, audiences, conversions, and budget management. Average Google Ads CPCs increased 42% between 2022 and 2025, which makes catching inefficiencies faster more valuable than ever. This guide covers the complete setup process, from Google Cloud Platform configuration to running your first automated workflow.
1,000+ Marketers Use Ryze
Automating hundreds of agencies
★★★★★4.9/5
What do you need to set up Google Ads MCP Server?
Setting up the Google Ads MCP Server requires several components: a Google Cloud Platform account for API access, a Google Ads account with developer permissions, Python 3.10+ for running the server locally, and Claude Desktop for the MCP client connection. The entire setup process takes 10-15 minutes if you have all prerequisites ready.
| Component | Requirement | Setup Time | Cost |
|---|---|---|---|
| Google Cloud Platform | Active GCP account with billing | 5 minutes | Free tier available |
| Google Ads Account | Developer token + API access | 2-3 minutes | Free |
| Python Environment | Python 3.10+ installed | 5-10 minutes | Free |
| Claude Desktop | Latest version with MCP support | 2 minutes | $20/month (Pro plan) |
Google Developer Token: You need a Google Ads Developer Token, which requires completing Google's API access application. This process typically takes 1-2 business days for approval. Test accounts get immediate access, but production tokens require a brief review process.
Alternative Setup Options: If you prefer a managed solution without the technical setup, Ryze AI provides a pre-configured MCP connector at get-ryze.ai/mcp that connects to Google Ads in under 2 minutes. For a broader comparison of Google Ads automation tools, see Top AI Tools for Google Ads Management in 2026.
How to install Google Ads MCP Server (6 steps)
This walkthrough covers the complete installation process using Google's official open-source implementation. Each step includes command-line instructions for macOS, Windows, and Linux environments. Total setup time: 15 minutes with all prerequisites ready.
Step 01
Configure Google Cloud Platform
Create a new Google Cloud project and enable the Google Ads API. Go to Google Cloud Console > Create Project > enter "Google Ads MCP" as project name > Create. Then navigate to APIs & Services > Library > search "Google Ads API" > Enable.
Step 02
Create OAuth 2.0 credentials
In Google Cloud Console, go to APIs & Services > Credentials > Create Credentials > OAuth 2.0 Client IDs. Select "Desktop application" as application type, name it "Google Ads MCP Client," and download the JSON file. Save it as client_secret.json in your project directory.
Step 03
Get Google Ads Developer Token
Log into your Google Ads account > Tools & Settings > API Center > request developer token. For testing, you can use your Google Ads customer ID immediately. For production access with live campaigns, Google reviews applications within 1-2 business days.
Step 04
Clone and install the MCP server
Clone Google's official repository and install dependencies. The server requires Python 3.10+ and several Google API libraries. Create a virtual environment to avoid conflicts with other Python projects.
Step 05
Configure environment variables
Create a .env file in your project root with your Google Ads developer token and OAuth configuration path. The server loads these automatically on startup and handles token refresh for continuous access.
Step 06
Test the connection
Run the MCP server to verify authentication works. The first run triggers an OAuth flow — follow the browser prompts to authorize access. Successful authentication creates a token.json file for automatic refresh.
You should see "MCP server starting on stdio" followed by authentication prompts. Complete the OAuth flow to establish the connection.
How do you connect Google Ads MCP Server to Claude Desktop?
Claude Desktop requires MCP server configuration through its JSON settings file. This file specifies the server path, execution environment, and connection parameters. The configuration enables Claude to launch the Google Ads MCP Server automatically and maintain persistent connections for real-time queries.
Locate your Claude Desktop configuration file:
Edit the configuration file and add your Google Ads MCP server. Replace the paths with your actual installation directory:
Restart Claude Desktop after saving the configuration. You should see a green MCP indicator when the connection is active. Test the integration by asking Claude: "Show me my Google Ads campaigns from last 7 days." If properly configured, Claude returns campaign performance data directly from your account.
Ryze AI — Autonomous Marketing
Skip the setup — let AI optimize your Google Ads 24/7
- ✓Automates Google, Meta + 5 more platforms
- ✓Handles your SEO end to end
- ✓Upgrades your website to convert better
2,000+
Marketers
$500M+
Ad spend
23
Countries
What Google Ads tasks can you automate with Claude MCP?
Once connected, the Google Ads MCP Server enables 7 core automation workflows that replace hours of manual analysis. Each workflow uses GAQL (Google Ads Query Language) to pull live data and Claude's reasoning to provide actionable insights. The average Google Ads account wastes 15-25% of budget on underperforming keywords and audiences — these workflows help identify and fix inefficiencies in real-time.
Workflow 01
Keyword Performance Analysis
Most Google Ads accounts have 20-30% of keywords driving 80% of conversions while the remaining keywords burn budget at 2-4x target CPA. Claude analyzes search term reports, identifies high-cost low-performing keywords, and calculates quality scores across your account. It flags keywords with CPA > 150% of target, low impression share due to budget constraints, and search terms that should be added as negative keywords.
Workflow 02
Search Term Mining
Search term reports reveal exactly what users typed before clicking your ads. Claude scans these reports for high-converting search terms not covered by your current keywords, identifies negative keyword opportunities to block irrelevant traffic, and suggests exact-match keywords to capture high-intent queries at lower CPCs. This workflow typically finds 10-15 new profitable keywords per month on active accounts.
Workflow 03
Bid Strategy Optimization
Google's automated bidding works well for established campaigns but often overspends on new keywords and audiences. Claude compares manual CPC performance against Target CPA and Maximize Conversions strategies, identifies campaigns where automated bidding is underperforming manual control, and recommends bid adjustments based on dayparting, device, and geographic performance data.
Workflow 04
Ad Copy Performance Analysis
RSAs (Responsive Search Ads) can have up to 15 headlines and 4 descriptions, creating 43,680 possible combinations. Claude analyzes which headline and description combinations drive the highest CTR and conversion rates, identifies underperforming ad variations, and suggests systematic A/B tests for creative optimization. Ad copy testing alone can improve CTR by 15-30% when done systematically.
Workflow 05
Audience Performance Review
Google Ads offers dozens of audience targeting options: demographics, interests, remarketing lists, similar audiences, and custom intent. Claude compares CPA and conversion volume across all audience segments, identifies high-performing audiences that should get increased budget allocation, and flags audiences with high spend but low conversion rates that are inflating overall account CPA.
Workflow 06
Budget Allocation Analysis
Most Google Ads accounts spread budget evenly across campaigns regardless of performance. Claude calculates marginal return on ad spend (ROAS) for each campaign to identify where additional budget would generate the highest incremental return. It simulates budget reallocation scenarios and projects the impact on total conversions and revenue. Smart budget allocation typically improves overall account ROAS by 10-20%.
Workflow 07
Conversion Tracking Audit
Broken conversion tracking is the silent killer of Google Ads performance. Claude verifies that conversion actions are firing correctly, identifies discrepancies between Google Ads conversion data and Google Analytics goals, and flags campaigns with suspiciously low conversion rates that may indicate tracking issues. Conversion tracking problems often go unnoticed for weeks, wasting thousands in optimization based on incomplete data.
How do you fix common Google Ads MCP Server issues?
Issue 1: "Authentication failed" error. This typically means your OAuth token expired or the client_secret.json file path is incorrect. Delete the existing token.json file and restart the server to trigger a fresh authentication flow. Verify the GOOGLE_ADS_OAUTH_CONFIG_PATH environment variable points to the correct JSON file.
Issue 2: "Developer token not approved" message. Google requires developer token approval for production accounts. Test accounts work immediately, but live campaigns need a 1-2 day review process. Apply for production access at ads.google.com > Tools > API Center if running queries against live campaigns.
Issue 3: Claude shows "MCP server not connected." Check your claude_desktop_config.json syntax — JSON is sensitive to trailing commas and quote matching. Restart Claude Desktop after any configuration changes. Verify the Python path points to your virtual environment's Python executable, not system Python.
Issue 4: "Permission denied" when querying campaigns. Your Google Ads account may not have API access enabled. Go to Google Ads > Tools & Settings > API Center > enable API access. Some enterprise accounts require admin approval for API connections.
Issue 5: Slow query performance. Large accounts with 10,000+ keywords can trigger API rate limits. Use date ranges in your queries to limit data volume. The Google Ads API has a 15,000 operations per day limit for standard access. For higher volume needs, request increased quotas through Google's developer console.
Sarah K.
Paid Media Manager
E-commerce Agency
We went from spending 10 hours a week on bid management to maybe 30 minutes reviewing Ryze’s recommendations. Our ROAS went from 2.4x to 4.1x in six weeks.”
4.1x
ROAS achieved
6 weeks
Time to result
95%
Less manual work
Frequently asked questions
Q: Is the Google Ads MCP Server free?
Yes, Google's open-source MCP server is free to use. You need Claude Pro ($20/month) for the MCP client connection. Google Ads API access is free with a developer token, though some enterprise features require Google Cloud billing.
Q: Can the MCP server make changes to my Google Ads account?
The current implementation provides read-only access for analysis and reporting. It cannot pause campaigns, adjust bids, or modify ads. For automated execution with safeguards, Ryze AI offers fully autonomous optimization.
Q: What is GAQL and how does Claude use it?
GAQL (Google Ads Query Language) is Google's SQL-like syntax for querying campaign data. Claude generates GAQL queries automatically based on your natural language questions, retrieves the data via MCP, and provides analysis.
Q: How long does Google Ads developer token approval take?
Test tokens are instant. Production tokens require 1-2 business days for Google's review. You can start development and testing immediately, then switch to production access once approved.
Q: Does this work with Google Ads manager accounts (MCCs)?
Yes, the MCP server supports manager accounts. Configure multiple customer IDs in your environment variables to analyze campaigns across different client accounts from a single Claude interface.
Q: What data can Claude access through the MCP server?
Claude can query campaigns, ad groups, keywords, ads, extensions, audiences, conversion actions, search terms, and performance metrics. It has the same data access as the Google Ads web interface.
Ryze AI — Autonomous Marketing
Connect Claude to Google Ads in under 2 minutes
- ✓Automates Google, Meta + 5 more platforms
- ✓Handles your SEO end to end
- ✓Upgrades your website to convert better
2,000+
Marketers
$500M+
Ad spend
23
Countries
