
In this article, we're going to show you to connect your Google ads account directly to Claude using MCP, so that you can chat with the data without having to even sign in to Google ads!
Think of MCP like Zapier, but for AI.
MCP (modular command protocol) is the middleman that connects AI assistants like Claude or ChatGPT directly to your data and services. It solves a fundamental limitation of AI assistants - they're great at understanding and generating text, but on their own, they can't actually do anything with your real-world tools.
Imagine asking Claude questions like...
- "Which campaigns had the highest ROI last month?"
- "Show me ads with declining CTR"
- "Analyze my search term performance across all campaigns"
This is what we're going to do!
Why MCP is a Game-Changer
In the past, AI assistants were limited to:
- Answering questions based on their training data
- Writing content like emails or reports
- Giving you instructions on how to do things
But with MCP, Claude can now:
- Directly access your Google Ads account data
- Analyze performance metrics in real-time
- Visualize trends across your campaigns
- Provide insights based on your actual advertising data
MCP creates a standardized way for Claude to communicate with external services like Google Ads, eliminating the complex "glue code" that developers previously needed to build for each integration.
What this means for you
Without getting technical, MCP means you can now have conversations with your Google Ads data through Claude. Instead of jumping between dashboards and downloading reports, you can simply ask questions like "Which campaigns had the highest CTR last month?" or "Show me ads with declining performance" and get immediate, intelligent responses.
This guide will walk you through setting up this powerful connection between Claude and your Google Ads account. Once configured, you'll have an AI marketing assistant that can analyze your advertising performance through natural conversation.
What You'll Need Before Starting
- Google Ads account with administrative access
- Claude Desktop app installed (download here)
- About 30-45 minutes of setup time
- Basic comfort with command line tools (don't worry, we'll guide you)

Step 1: Google Cloud & API Setup (10 minutes)
1.1: Create a Google Cloud Project
- Go to Google Cloud Console
- Click "Select a project" in the top menu bar
- Click "NEW PROJECT" in the popup window

This step requires waiting for Google's approval, so start it early!
2.1: Access the API Center
- Sign in to your Google Ads account at ads.google.com
- Click the wrench icon (Tools & Settings) in the top right corner
- Under "Setup", click "API Center"

3. Click "ENABLE" button
1.3: Create OAuth Credentials
- In the left navigation menu, click "APIs & Services" then "Credentials"
- Click "+ CREATE CREDENTIALS" at the top, then "OAuth client ID" Show Image
- For Application type, select "Desktop app"
- Name your OAuth client (e.g., "Claude Desktop Integration")
- Click "CREATE"
- Download the JSON file (it will be named something like client_secret_XXXX.json)
- Save this file somewhere you can easily find it later (we'll need it in Step 4)

Step 2: Get Your Google Ads Developer Token (1-3 business days)
This step requires waiting for Google's approval, so start it early!
2.1: Access the API Center
- Sign in to your Google Ads account at ads.google.com
- Click the wrench icon (Tools & Settings) in the top right corner
- Under "Setup", click "API Center"
2.2: Apply for a Developer Token
- Accept the Terms of Service if prompted
- Click "APPLY FOR TOKEN"
- Fill out the application form:
- Explain you're using it with Claude AI for analyzing your own personal Google ads account.
- Mention you're using the MCP integration for natural language analysis
- Submit and wait for approval (usually takes 1-3 business days)
- Once approved, you'll see your developer token in the API Center
- Save this token somewhere secure - you'll need it in Step 4
Note: You'll start with a test token that has some limitations but is sufficient for personal use.
Step 3: Install Required Software (15 minutes)
3.1: Install Python
- Download Python (version 3.11 or newer):
- Windows/Mac: python.org
- During installation:
- ✅ IMPORTANT: Check "Add Python to PATH"
- Complete the installation with default settings
- Verify installation:
- Open Terminal (Mac) or Command Prompt (Windows)
- Type python --version and press Enter
- You should see a version number like Python 3.11.x
3.2: Install Node.js
- Download Node.js from nodejs.org
- Run the installer with default settings
- Verify installation:
- In Terminal/Command Prompt, type node --version
- You should see a version number like v18.x.x
3.3: Ensure Claude Desktop is Installed
If you haven't already, download and install Claude Desktop app from anthropic.com
Step 4: Download & Configure the Google Ads MCP Tool (10 minutes)
4.1: Download the Tool
- Go to https://github.com/cohnen/mcp-google-ads
- Click the green "Code" button
- Select "Download ZIP" Show Image
- Find the downloaded ZIP file and extract it to a location you can easily find
- Recommendation: Extract to a folder called google-ads-claude in your Documents folder
4.2: Open Terminal/Command Prompt
- Windows: Search for "Command Prompt" in the start menu
- Mac: Search for "Terminal" in Spotlight (press Cmd+Space and type "Terminal")
4.3: Navigate to the Tool Folder
Type cd followed by the path to where you extracted the files. For example:
- Windows: cd C:\Users\YourName\Documents\google-ads-claude\mcp-google-ads-main
- Mac: cd ~/Documents/google-ads-claude/mcp-google-ads-main
Some coding bits next...
4.4: Create a Virtual Environment
In your terminal, run these commands one by one:
bash
# Create virtual environment python -m venv .venv # Activate the environment # For Windows: .venv\Scripts\activate # For Mac: source .venv/bin/activate
You'll know it worked when you see (.venv) at the beginning of your command line.
4.5: Install Required Packages
Run this command:
bash
pip install -r requirements.txt
If you encounter any issues, try:
bash
pip install mcp
Step 5: Connect Claude to Google Ads (10 minutes)
5.1: Find Your Claude Configuration File
- Open Terminal/Command Prompt again
Enter the command to edit Claude's configuration:
Mac:
nano ~/Library/Application\ Support/Claude/claude_desktop_config.json
Windows:
notepad %APPDATA%\Claude\claude_desktop_config.json
This will open a text editor with Claude's configuration file.
Few more coding bits...
5.2: Add Google Ads Configuration
- If the file is empty, add the following code. If it contains existing configuration, carefully add the googleAdsServer section to the mcpServers object:
json
{ "mcpServers": { "googleAdsServer": { "command": "PYTHON_PATH", "args": ["SERVER_SCRIPT_PATH"], "env": { "GOOGLE_ADS_AUTH_TYPE": "oauth", "GOOGLE_ADS_CREDENTIALS_PATH": "CREDENTIALS_PATH", "GOOGLE_ADS_DEVELOPER_TOKEN": "YOUR_DEVELOPER_TOKEN" } } } }
- Replace the placeholder values:
- PYTHON_PATH: Path to Python in your virtual environment
- Windows: C:\\Users\\YourName\\Documents\\google-ads-claude\\mcp-google-ads-main\\.venv\\Scripts\\python.exe
- Mac: /Users/YourName/Documents/google-ads-claude/mcp-google-ads-main/.venv/bin/python
- SERVER_SCRIPT_PATH: Path to the server script
- Windows: C:\\Users\\YourName\\Documents\\google-ads-claude\\mcp-google-ads-main\\google_ads_server.py
- Mac: /Users/YourName/Documents/google-ads-claude/mcp-google-ads-main/google_ads_server.py
- CREDENTIALS_PATH: Path to your OAuth credentials file from Step 1.3
- Windows: C:\\Users\\YourName\\Downloads\\client_secret_XXXX.json
- Mac: /Users/YourName/Downloads/client_secret_XXXX.json
- YOUR_DEVELOPER_TOKEN: Your Google Ads API Developer Token from Step 2.2
- PYTHON_PATH: Path to Python in your virtual environment
⚠️ Important: On Windows, you must use double backslashes (\\) in file paths within this JSON file.
5.3: Example Configuration
Here's an example of a complete configuration file:
json
{ "mcpServers": { "googleAdsServer": { "command": "C:\\Users\\JaneDoe\\Documents\\google-ads-claude\\mcp-google-ads-main\\.venv\\Scripts\\python.exe", "args": ["C:\\Users\\JaneDoe\\Documents\\google-ads-claude\\mcp-google-ads-main\\google_ads_server.py"], "env": { "GOOGLE_ADS_AUTH_TYPE": "oauth", "GOOGLE_ADS_CREDENTIALS_PATH": "C:\\Users\\JaneDoe\\Downloads\\client_secret_123456789.json", "GOOGLE_ADS_DEVELOPER_TOKEN": "AbCd1234XyZ_9876", "GOOGLE_ADS_LOGIN_CUSTOMER_ID": "123-456-7890" } } } }
Note: The GOOGLE_ADS_LOGIN_CUSTOMER_ID is optional. Include it only if you're using a Google Ads Manager Account (format: XXX-XXX-XXXX).
5.4: Save the Configuration File
- Mac: Press Ctrl+O, then Enter, then Ctrl+X to exit
- Windows: Click File > Save, then close Notepad
5.5: Restart Claude Desktop
Close and reopen Claude Desktop app to apply your changes.
Step 6: Start Using Claude with Google Ads!
Now you're finally ready to start playing around!
Open Claude Desktop - If everything is set up correctly, you'll be prompted to authorize Claude to access your Google Ads data the first time you try to use it.
Here's some prompt ideas that you can feed into Claude now that your Google ads data is connected:

Having trouble?
If you're having trouble getting this all connected and set up, or you would rather a solution that does this all for you, don't worry, we've got you covered.
Just go to PPC.io and sign up to our tool - We've integrated literally every single AI tool within our interface and completed all of the heavy coding, so you can just sign up and go!
Troubleshooting
Google Authentication Failed
- Verify your credentials file path is correct
- Check if your OAuth client is properly configured
- Try downloading a fresh credentials file
Developer Token Issues
- Ensure you've copied the token correctly
- Verify the token is active in your Google Ads API Center
- Remember that test tokens have usage limitations
Claude Can't Find the Tool
- Check that paths in the configuration file are correct
- Make sure to use double backslashes (\\) in Windows paths
- Verify the virtual environment is in the right location
Command Not Found Errors
- Ensure Python and Node.js are added to your system PATH
- Try reinstalling Python with the "Add to PATH" option checked
- Verify the virtual environment was created correctly
Advanced Features
Once you're comfortable with the basic setup, you can try:
- Custom GAQL queries for specific data analysis
- Automated weekly reports through Claude
- Performance visualizations and trend analysis
- Budget optimization recommendations