MCP
How to Set Up MCP Server for Google Ads API — Complete 2026 Developer Guide
Setting up MCP server for Google Ads API connects Claude directly to your advertising data for real-time campaign analysis. This guide covers authentication, configuration, deployment options, and 12 automated workflows that cut account management time from 20 hours to under 3 hours weekly.
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 MCP server for Google Ads API?
An MCP (Model Context Protocol) server for Google Ads API is a bridge that allows AI models like Claude to directly access your Google Ads account data without manual exports or copy-pasting. When you set up MCP server for Google Ads API, Claude can pull campaign metrics, analyze keyword performance, detect bid anomalies, and generate optimization reports in real-time — all through natural language queries.
The MCP server acts as an intermediary that handles authentication, API calls, and data formatting. Instead of downloading CSV reports from Google Ads Manager and uploading them to Claude, you ask questions like "Which campaigns have CTR < 2% this month?" and get instant, structured answers backed by live account data. Google Ads API serves over 2.8 billion daily requests, but most marketers still rely on manual reporting that's hours or days behind real performance.
This guide covers three setup methods: the official Google MCP server (15-20 minutes), Ryze AI's managed connector (2 minutes), and self-hosted deployment on Google Cloud Run (30-45 minutes). Each method has different tradeoffs between setup complexity, control, and ongoing maintenance. For broader context on Claude's capabilities with Google Ads, see Claude Skills for Google Ads and How to Use Claude for Google Ads.
1,000+ Marketers Use Ryze
Automating hundreds of agencies
★★★★★4.9/5
What are the 3 ways to set up MCP server for Google Ads API?
There are three primary methods to connect Claude to Google Ads through MCP, each optimized for different user types and requirements. The right choice depends on your technical expertise, control needs, and how frequently you require data updates.
| Method | Setup Time | Technical Level | Best For |
|---|---|---|---|
| Ryze AI Connector | 2 minutes | None required | Marketers who want instant setup |
| Official Google MCP | 15-20 minutes | Basic development | Developers wanting direct integration |
| Cloud Run Deployment | 30-45 minutes | Advanced development | Teams needing scalable, shared access |
Method 1: Ryze AI Connector provides the fastest path to MCP connectivity. Visit get-ryze.ai/mcp, authenticate with your Google account, and receive a pre-configured MCP endpoint. No credential management, no YAML files, no troubleshooting — just instant access to your Google Ads data.
Method 2: Official Google MCP Server requires Python 3.12+, a Google Ads developer token, OAuth credentials, and a google-ads.yaml configuration file. You install the server locally using pipx and connect Claude Desktop through MCP settings. This method gives you full control over the implementation and follows Google's official standards.
Method 3: Cloud Run Deployment hosts the MCP server on Google Cloud infrastructure. This approach supports multiple team members accessing the same MCP endpoint, handles scaling automatically, and provides enterprise-grade security. It requires Google Cloud project setup, Docker containerization, and network configuration.
What are the prerequisites for setting up MCP server for Google Ads API?
Before configuring your MCP server for Google Ads API, ensure you have the necessary accounts, credentials, and system requirements. Missing any of these components will cause authentication failures or connection errors during setup.
System Requirements
- •Python 3.12 or higher with pip installed
- •Claude Desktop application (latest version)
- •Node.js 18+ (for Ryze AI connector method)
- •Git command-line tools
- •Google Cloud CLI (for Cloud Run deployment)
Google Ads Account Requirements
- •Active Google Ads account with administrative access
- •Google Ads Developer Token (basic access minimum)
- •Manager Account (MCC) access for multi-account setups
- •API access enabled in Google Ads settings
Authentication Credentials
- •OAuth 2.0 Client ID and Client Secret
- •OAuth Refresh Token (generated during initial setup)
- •Login Customer ID (your Google Ads customer ID)
- •Google Cloud Project (for Cloud Run deployment)
The Google Ads Developer Token requires approval from Google, which typically takes 1-3 business days for legitimate use cases. Specify "MCP integration for automated campaign analysis" as your intended usage to expedite approval. Standard access tokens allow read-only operations, while test tokens work only with test accounts.
How do you set up MCP server for Google Ads API step by step?
This walkthrough covers the official Google MCP server installation, which provides direct integration without third-party dependencies. The entire process takes 15-20 minutes for users with existing Google Cloud accounts, or 30-45 minutes for first-time setups.
Step 01
Obtain Google Ads Developer Token
Navigate to Google Ads > Tools & Settings > API Center. Click "Apply for token" and complete the application with your intended use case. Specify "MCP integration for automated campaign analysis and reporting" as your usage description. Include details about data handling frequency and technical implementation. Developer token approval typically takes 1-3 business days for legitimate applications.
Step 02
Configure OAuth 2.0 Credentials
In Google Cloud Console, create a new project or select an existing one. Navigate to APIs & Services > Credentials > Create Credentials > OAuth 2.0 Client ID. Choose "Desktop application" as the application type. Download the credentials JSON file and note your Client ID and Client Secret for the configuration file.
Step 03
Generate OAuth Refresh Token
Use Google's OAuth Playground or the google-ads-python authentication script to generate your refresh token. In OAuth Playground, configure your Client ID and Secret, authorize the Google Ads API scope, and exchange the authorization code for an access token and refresh token. Save the refresh token securely — it doesn't expire and provides ongoing API access.
Step 04
Install the MCP Server
Install Python 3.12+ and pipx if not already available. Run the installation command to download and configure the Google Ads MCP server. The installation pulls the latest version from Google's official repository and sets up all required dependencies automatically.
Step 05
Configure Claude Desktop
Open Claude Desktop > Settings > MCP Servers > Add Server. Paste the configuration JSON with your environment variables. Claude Desktop will validate the connection and show a green indicator when the MCP server is active and authenticated successfully.
Step 06
Test the Connection
Restart Claude Desktop to load the new MCP configuration. Ask Claude: "What Google Ads customers do I have access to?" If configured correctly, Claude returns a list of your accessible accounts with customer IDs and descriptive names. Test additional functionality by requesting campaign performance data for the last 30 days.
How do you configure authentication and credentials?
The MCP server for Google Ads API requires a google-ads.yaml configuration file containing your authentication credentials. This file handles token refresh automatically and maintains secure access to your advertising data without manual re-authentication.
Creating the google-ads.yaml File
Create a file named google-ads.yaml in your home directory (~/google-ads.yaml on macOS/Linux, C:\Users\YourName\google-ads.yaml on Windows). This file contains five required credentials that authenticate your MCP server with the Google Ads API.
Environment Variables
The MCP server supports environment variables for credential management in production environments. Set GOOGLE_ADS_CREDENTIALS to specify an alternate path for your configuration file, or use USE_GOOGLE_OAUTH_ACCESS_TOKEN for temporary access token authentication.
Security Best Practices
Store credentials securely and limit file permissions. On Unix systems, set google-ads.yaml permissions to 600 (read/write for owner only). Never commit credentials to version control. For production deployments, use Google Secret Manager or equivalent secure credential storage.
Ryze AI — Autonomous Marketing
Skip the setup — connect Claude to Google Ads in 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
What automation workflows can you run with MCP server for Google Ads API?
Once your MCP server is configured, you can automate 12 essential workflows that traditionally require hours of manual analysis. These workflows cover account auditing, performance optimization, budget management, and competitive analysis. The average Google Ads account wastes 15-25% of its budget on underperforming keywords and ad groups.
Workflow 01
Bid Anomaly Detection
Google Ads automated bidding can malfunction, leading to bid spikes that inflate CPCs by 50-200%. Claude analyzes bid changes across keywords, identifies statistical outliers, and flags bidding anomalies that require immediate attention. This workflow catches bid inflation within hours instead of weeks.
Workflow 02
Quality Score Monitoring
Quality Score directly impacts ad position and cost-per-click. A drop from 8 to 6 can increase CPCs by 25%. Claude tracks Quality Score changes across keywords, identifies factors causing score degradation, and provides specific improvement recommendations for landing page, ad relevance, and expected CTR.
Workflow 03
Search Term Analysis
Search term reports reveal which queries actually trigger your ads. Claude identifies high-volume search terms with poor conversion rates, suggests negative keywords to add, and finds untapped opportunities for exact-match keywords. Proper search term optimization typically reduces wasted spend by 20-30%.
Workflow 04
Budget Pacing Analysis
Campaigns that exhaust budgets early lose impression share and conversions. Claude calculates monthly budget pacing, identifies campaigns running ahead or behind schedule, and recommends budget adjustments. Proper pacing ensures consistent visibility throughout the month.
Workflow 05
Ad Copy Performance Ranking
Within each ad group, different ad copies perform vastly differently. Claude ranks all ads by CTR and conversion rate, identifies top and bottom performers, and suggests which ads to pause or promote. Regular ad rotation testing can improve CTR by 15-40%.
Workflow 06
Impression Share Analysis
Lost impression share indicates missed opportunities. Claude analyzes impression share loss due to budget vs. ad rank, calculates potential additional conversions, and prioritizes which campaigns need budget increases vs. bid optimizations.
Workflows 07-12
Additional Automation Capabilities
The remaining workflows include device performance analysis (mobile vs. desktop conversion rates), dayparting optimization (identifying peak performance hours), geographic performance auditing (state/city-level ROAS analysis), competitor auction insights (shared budget analysis), landing page performance correlation, and automated reporting generation. Each workflow follows the same pattern: Claude pulls live data, identifies optimization opportunities, and provides actionable recommendations.
Sarah K.
Paid Media Manager
E-commerce Agency
The MCP setup took 20 minutes and completely changed how we manage Google Ads. We catch bid anomalies the same day instead of weeks later. Our Quality Scores improved 23% in two months.”
23%
QS improvement
2 months
Time to result
20 min
Setup time
What are common issues when setting up MCP server for Google Ads API?
Issue 1: Authentication failures. The most common problem is incorrect credentials in google-ads.yaml. Verify your developer token is approved (not still pending), refresh token hasn't expired, and client ID/secret match your OAuth application. Test credentials using Google's authentication example script before configuring MCP.
Issue 2: MCP server connection timeout. Claude shows "Server not responding" when the MCP server crashes or can't authenticate. Check Claude Desktop logs (Help > Show Logs) for error details. Common causes include missing Python dependencies, incorrect file paths, or network connectivity issues.
Issue 3: Permission denied errors. API calls fail with "User doesn't have permission to access customer" when your developer token lacks access to specific accounts. Verify your Google account has administrative access to all customer accounts you're trying to query.
Issue 4: Python version incompatibility. The official MCP server requires Python 3.12+. Running on older Python versions causes import errors or cryptographic failures. Use `python --version` to verify your installation and upgrade if necessary.
Issue 5: Rate limiting. Excessive API calls trigger Google's rate limits, causing temporary failures. The MCP server should handle rate limiting automatically, but large data requests may need pagination. If you encounter rate limits, reduce query scope or add delays between requests.
Frequently asked questions
Q: How long does it take to set up MCP server for Google Ads API?
Setup takes 15-20 minutes with existing Google Cloud accounts, or 30-45 minutes for first-time configurations. The Ryze AI connector reduces this to under 2 minutes with managed authentication.
Q: Do I need a Google Ads developer token?
Yes, a Google Ads developer token is required for API access. Apply through Google Ads > Tools & Settings > API Center. Approval typically takes 1-3 business days for legitimate use cases.
Q: Can the MCP server make changes to my Google Ads account?
By default, the MCP server operates in read-only mode. Enable mutations by setting ADS_MCP_ENABLE_MUTATIONS=true, but use caution as this allows Claude to modify campaigns, keywords, and bids.
Q: Is MCP server for Google Ads API free to use?
The official Google MCP server is open source and free. You need Claude Pro ($20/month) to use MCP functionality. Google Ads API usage is free within rate limits and quotas.
Q: What data can I access through MCP server?
The MCP server provides access to campaign metrics, keyword performance, ad copy data, search terms, Quality Scores, auction insights, and account settings. All data available through Google Ads API is accessible.
Q: How does this compare to manual reporting?
Manual reporting typically takes 3-5 hours weekly and uses stale data. MCP automation delivers the same insights in minutes with real-time data, plus automated anomaly detection and optimization recommendations.
Ryze AI — Autonomous Marketing
Connect Claude to Google Ads without the technical setup
- ✓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
