Google Ads MCP Server
Enables programmatic management of Google Ads campaigns, allowing users to monitor performance metrics, update budgets, and toggle campaign statuses. It supports real-time analytics, top performer analysis, and reporting in CSV or JSON formats.
README
Google Ads MCP Server (Node.js)
A Model Context Protocol (MCP) server for managing Google Ads campaigns programmatically, built with Node.js.
Features
Campaign Management
- ✅ Fetch campaign performance data
- ✅ Get detailed campaign information
- ✅ Pause/enable campaigns
- ✅ Update campaign budgets
Analytics & Reporting
- ✅ Performance summaries
- ✅ Top performer analysis
- ✅ Export to CSV/JSON
- ✅ Custom date ranges
Real-time Data
- ✅ Live API integration
- ✅ Up-to-date metrics
- ✅ Conversion tracking
- ✅ Budget monitoring
Installation
Global Installation (Recommended for CLI use)
npm install -g @samihalawa/google-ads-mcp-server
Local Installation (For project integration)
npm install @samihalawa/google-ads-mcp-server
Or with pnpm:
pnpm add @samihalawa/google-ads-mcp-server
Configuration
Simple .env Configuration (No YAML file needed!)
Create a .env file with your Google Ads API credentials as inline JSON:
# Google Ads API Configuration (JSON format - all in one line)
GOOGLE_ADS_CONFIG='{"client_id":"YOUR_CLIENT_ID.apps.googleusercontent.com","client_secret":"YOUR_CLIENT_SECRET","developer_token":"YOUR_DEVELOPER_TOKEN","refresh_token":"YOUR_REFRESH_TOKEN","login_customer_id":"YOUR_MANAGER_CUSTOMER_ID"}'
# Customer ID to query (without dashes)
GOOGLE_ADS_CUSTOMER_ID=1234567890
That's it! No separate YAML files needed. Everything is in your .env file.
Example Configuration
GOOGLE_ADS_CONFIG='{"client_id":"963208150325-mmhibhl91g39ma9jsvrgacpleraq4nfu.apps.googleusercontent.com","client_secret":"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","developer_token":"i525AeFTAacFOtQtWBjY6g","refresh_token":"1//04OmKZJ58yhQaCgYIARAAGAQSNwF-L9IrfyrhE7W2zk00iStBE8dCRazdeUgXiMVxH-WIr9PEh6W3_RvjRKSZx-FH3l3Dun5vWOc","login_customer_id":"4850172260"}'
GOOGLE_ADS_CUSTOMER_ID=1248495560
Usage
Running the Server
# Load .env and run
node server.js
Or with npx (no installation):
npx @samihalawa/google-ads-mcp-server
Using with MCP Clients
Add to your MCP client configuration (e.g., Claude Desktop):
{
"mcpServers": {
"google-ads": {
"command": "npx",
"args": ["@samihalawa/google-ads-mcp-server"],
"env": {
"GOOGLE_ADS_CONFIG": "{\"client_id\":\"YOUR_CLIENT_ID\",\"client_secret\":\"YOUR_SECRET\",\"developer_token\":\"YOUR_TOKEN\",\"refresh_token\":\"YOUR_REFRESH\",\"login_customer_id\":\"YOUR_MANAGER_ID\"}",
"GOOGLE_ADS_CUSTOMER_ID": "1234567890"
}
}
}
}
Using with manus-mcp-cli
# Set environment variables
export GOOGLE_ADS_CONFIG='{"client_id":"...","client_secret":"...","developer_token":"...","refresh_token":"...","login_customer_id":"..."}'
export GOOGLE_ADS_CUSTOMER_ID="1234567890"
# List available tools
manus-mcp-cli tool list --server google-ads
# Get campaigns
manus-mcp-cli tool call get_campaigns --server google-ads --input '{"days": 30}'
# Get performance summary
manus-mcp-cli tool call get_performance_summary --server google-ads --input '{"days": 7}'
# Pause a campaign
manus-mcp-cli tool call pause_campaign --server google-ads --input '{"campaign_id": "23207843655"}'
Available Tools
1. get_campaigns
Fetch all campaigns with performance metrics.
Parameters:
days(number, optional): Number of days to look back (default: 30)status(string, optional): Filter by status - ENABLED, PAUSED, REMOVED, or ALL (default: ENABLED)
Example:
{
"days": 30,
"status": "ENABLED"
}
2. get_campaign_details
Get detailed information about a specific campaign.
Parameters:
campaign_id(string, required): The campaign IDdays(number, optional): Number of days to look back (default: 30)
Example:
{
"campaign_id": "23207843655",
"days": 30
}
3. get_performance_summary
Get overall account performance summary.
Parameters:
days(number, optional): Number of days to look back (default: 30)
Example:
{
"days": 7
}
4. get_top_performers
Get top performing campaigns by specified metric.
Parameters:
metric(string, optional): Metric to rank by - ctr, conversions, cost, clicks, impressions (default: ctr)limit(number, optional): Number of top campaigns to return (default: 5)days(number, optional): Number of days to look back (default: 30)
Example:
{
"metric": "ctr",
"limit": 5,
"days": 30
}
5. pause_campaign
Pause a specific campaign.
Parameters:
campaign_id(string, required): The campaign ID to pause
Example:
{
"campaign_id": "23207843655"
}
6. enable_campaign
Enable/resume a paused campaign.
Parameters:
campaign_id(string, required): The campaign ID to enable
Example:
{
"campaign_id": "23207843655"
}
7. update_campaign_budget
Update the daily budget for a campaign.
Parameters:
campaign_id(string, required): The campaign IDbudget_euros(number, required): New daily budget in euros
Example:
{
"campaign_id": "23207843655",
"budget_euros": 20.00
}
8. export_report
Export campaign data to CSV or JSON format.
Parameters:
format(string, required): Export format - csv or jsondays(number, optional): Number of days to look back (default: 30)
Example:
{
"format": "csv",
"days": 30
}
Example Workflows
Daily Campaign Monitoring
# Get performance summary
manus-mcp-cli tool call get_performance_summary --server google-ads --input '{"days": 1}'
# Check top performers
manus-mcp-cli tool call get_top_performers --server google-ads --input '{"metric": "conversions", "limit": 3, "days": 7}'
Campaign Optimization
# Get campaign details
manus-mcp-cli tool call get_campaign_details --server google-ads --input '{"campaign_id": "23207843655", "days": 30}'
# Update budget if performing well
manus-mcp-cli tool call update_campaign_budget --server google-ads --input '{"campaign_id": "23207843655", "budget_euros": 25.00}'
# Pause if underperforming
manus-mcp-cli tool call pause_campaign --server google-ads --input '{"campaign_id": "23207843655"}'
Reporting
# Export to CSV
manus-mcp-cli tool call export_report --server google-ads --input '{"format": "csv", "days": 30}'
# Export to JSON
manus-mcp-cli tool call export_report --server google-ads --input '{"format": "json", "days": 7}'
Troubleshooting
"GOOGLE_ADS_CONFIG environment variable is required"
- Make sure you've set the
GOOGLE_ADS_CONFIGenvironment variable - Check that the JSON is valid and properly escaped
- Ensure all required fields are present
"Failed to initialize Google Ads client"
- Verify all credentials are correct
- Ensure refresh token is still valid
- Check that the JSON format is correct
"Campaign not found"
- Verify the campaign ID is correct
- Check that you have access to the campaign
- Ensure the campaign hasn't been removed
"Unauthorized" errors
- Refresh token may have expired - generate a new one
- Check that the developer token is approved
- Verify OAuth credentials are correct
API Rate Limits
Google Ads API has rate limits:
- Basic access: 15,000 operations per day
- Standard access: 40,000 operations per day
The MCP server automatically handles rate limiting and retries.
Security
- Never commit
.envto version control - Store credentials securely
- Use environment variables for sensitive data
- Rotate refresh tokens regularly
- The
.gitignorefile already excludes.envfiles
Support
For issues or questions:
- Check the Google Ads API documentation
- Review the MCP specification
- Check server logs for error messages
- Open an issue on GitHub
License
MIT License - See LICENSE file for details
Version History
1.1.0 (2025-11-24)
- Breaking Change: Switched from YAML to inline JSON configuration in .env
- Removed js-yaml dependency
- Simplified configuration - no separate files needed
- Updated documentation
1.0.0 (2025-11-24)
- Initial Node.js release
- 8 core tools for campaign management
- Real-time API integration
- CSV/JSON export support
- Complete documentation
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。