mcp-gads
Enables querying and managing Google Ads campaigns, keywords, assets, and more via natural language, with support for multiple MCP clients.
README
<p align="center"> <img src="logo.svg" alt="mcp-gads" width="280" /> </p>
<p align="center"> Google Ads MCP server — query campaigns, keywords, assets & more via natural language. <br/> Built with Bun + TypeScript. Works with Claude, Cursor, and any MCP client. </p>
Quick Start
1. Get Credentials
You need a Google Ads API developer token and OAuth client credentials.
- Download your OAuth client JSON from Google Cloud Console
- Set environment variables:
export GOOGLE_ADS_DEVELOPER_TOKEN=your-token
export GOOGLE_ADS_CREDENTIALS_PATH=./credentials.json
- Run the setup helper to authorize:
npx mcp-gads setup
This opens your browser, completes OAuth, and saves a refresh token.
2. Add to Claude Code
claude mcp add google-ads --scope user --transport stdio \
-e GOOGLE_ADS_DEVELOPER_TOKEN=your-token \
-e GOOGLE_ADS_CREDENTIALS_PATH=/path/to/credentials.json \
-- npx -y mcp-gads@latest
That's it. Restart Claude Code and the tools are available. Every session runs the latest version automatically.
Also works with
bunx mcp-gads@latestif you have Bun. Requires Node 18+ when running vianpx.
If your environment blocks npm registry access at runtime, install once and run the published binary name directly:
npm i -g mcp-gads@latest
claude mcp add google-ads --scope user --transport stdio \
-e GOOGLE_ADS_DEVELOPER_TOKEN=your-token \
-e GOOGLE_ADS_CREDENTIALS_PATH=/path/to/credentials.json \
-- mcp-gads
<details> <summary>Alternative: standalone binary</summary>
Download a pre-built binary from Releases:
| Platform | File |
|---|---|
| macOS (Apple Silicon) | mcp-gads-darwin-arm64 |
| macOS (Intel) | mcp-gads-darwin-x64 |
| Linux | mcp-gads-linux-x64 |
| Windows | mcp-gads-windows-x64.exe |
macOS / Linux:
curl -Lo mcp-gads https://github.com/pijusz/mcp-gads/releases/latest/download/mcp-gads-darwin-arm64
chmod +x mcp-gads
sudo mv mcp-gads /usr/local/bin/
claude mcp add google-ads --scope user --transport stdio \
-e GOOGLE_ADS_DEVELOPER_TOKEN=your-token \
-e GOOGLE_ADS_CREDENTIALS_PATH=/path/to/credentials.json \
-- /usr/local/bin/mcp-gads
Windows (PowerShell):
Invoke-WebRequest -Uri "https://github.com/pijusz/mcp-gads/releases/latest/download/mcp-gads-windows-x64.exe" -OutFile "$env:LOCALAPPDATA\mcp-gads.exe"
claude mcp add google-ads --scope user --transport stdio -e GOOGLE_ADS_DEVELOPER_TOKEN=your-token -e GOOGLE_ADS_CREDENTIALS_PATH=C:\path\to\credentials.json -- "%LOCALAPPDATA%\mcp-gads.exe"
</details>
ChatGPT Codex
Codex uses TOML, not JSON. Install once, then add to ~/.codex/config.toml:
npm i -g mcp-gads
[mcp_servers.gads]
command = "mcp-gads"
[mcp_servers.gads.env]
GOOGLE_ADS_DEVELOPER_TOKEN = "your-token"
GOOGLE_ADS_CREDENTIALS_PATH = "/absolute/path/to/credentials.json"
Three gotchas that cause silent failures on Codex:
- Don't use
npx -ywithout raising the timeout. Codex's defaultstartup_timeout_secis 10s, which is too short for npx's first-run download. A global install (above) or the prebuilt binary sidesteps this entirely. If you must use npx, addstartup_timeout_sec = 30. - Env vars must go under
[mcp_servers.gads.env]. Codex does not inherit the parent shell environment into stdio servers — exporting vars in your shell won't reach the server. - Use absolute paths for
GOOGLE_ADS_CREDENTIALS_PATH. Codex spawns the server with its own cwd, so relative paths silently miss.
On Windows some Codex versions use startup_timeout_ms = 20000 instead of _sec.
Claude Desktop
Add to your claude_desktop_config.json:
<details> <summary>Using npx (auto-updates)</summary>
{
"mcpServers": {
"google-ads": {
"command": "npx",
"args": ["-y", "mcp-gads@latest"],
"env": {
"GOOGLE_ADS_DEVELOPER_TOKEN": "your-token",
"GOOGLE_ADS_CREDENTIALS_PATH": "/path/to/credentials.json"
}
}
}
}
</details>
<details> <summary>Using binary (macOS / Linux)</summary>
{
"mcpServers": {
"google-ads": {
"command": "/usr/local/bin/mcp-gads",
"env": {
"GOOGLE_ADS_DEVELOPER_TOKEN": "your-token",
"GOOGLE_ADS_CREDENTIALS_PATH": "/path/to/credentials.json"
}
}
}
}
</details>
<details> <summary>Using binary (Windows)</summary>
{
"mcpServers": {
"google-ads": {
"command": "C:\\Users\\YOU\\AppData\\Local\\mcp-gads.exe",
"env": {
"GOOGLE_ADS_DEVELOPER_TOKEN": "your-token",
"GOOGLE_ADS_CREDENTIALS_PATH": "C:\\path\\to\\credentials.json"
}
}
}
}
</details>
Tools (39)
Account Management
| Tool | Description |
|---|---|
list_accounts |
List all accessible Google Ads accounts |
get_account_currency |
Get the currency code for an account |
get_account_hierarchy |
Get MCC account tree (manager -> client) |
Queries
| Tool | Description |
|---|---|
execute_gaql_query |
Run any GAQL query (table output) |
run_gaql |
Run GAQL with format options (table/json/csv) |
get_gaql_help |
GAQL reference guide with syntax, resources, and examples |
list_resources |
List valid GAQL FROM clause resources |
Campaigns
| Tool | Description |
|---|---|
get_campaign_performance |
Campaign metrics (impressions, clicks, cost, conversions) |
get_budget_utilization |
Budget amounts vs actual spend |
Ads
| Tool | Description |
|---|---|
get_ad_performance |
Ad-level performance metrics |
get_ad_creatives |
RSA headlines, descriptions, final URLs |
Assets
| Tool | Description |
|---|---|
get_image_assets |
List image assets with URLs and dimensions |
download_image_asset |
Download a specific image asset to disk |
get_asset_usage |
Find where assets are used (campaigns, ad groups) |
analyze_image_assets |
Image asset performance with metrics |
Keywords
| Tool | Description |
|---|---|
generate_keyword_ideas |
Keyword Planner suggestions from seed keywords |
get_keyword_volumes |
Historical search volume for specific keywords |
get_quality_scores |
Quality scores with component breakdown |
get_search_terms |
Actual search queries triggering your ads |
Geographic & Device
| Tool | Description |
|---|---|
get_geographic_performance |
Performance by location |
get_device_performance |
Performance by device type |
Insights
| Tool | Description |
|---|---|
get_recommendations |
Google's AI optimization suggestions |
get_change_history |
Recent account changes |
Extended Tools (disabled by default)
Enable with GOOGLE_ADS_ENABLE_EXTENDED_TOOLS=true:
| Tool | Description |
|---|---|
get_ad_group_performance |
Ad group metrics with optional campaign filter |
get_conversion_actions |
Conversion actions with settings and performance |
get_account_summary |
Quick dashboard: totals + top 5 campaigns |
get_impression_share |
Competitive position: impression share and lost IS |
get_ad_schedule_performance |
Performance by hour or day of week |
get_audience_performance |
Demographics: age range and gender breakdowns |
get_landing_page_performance |
Landing page URLs with metrics |
get_placement_performance |
Display/PMax placement details |
get_asset_group_performance |
PMax asset group metrics and ad strength |
get_video_performance |
YouTube/video view rates and quartile completion |
get_labels |
Labels and their campaign/ad group assignments |
Write Tools (disabled by default)
Enable with GOOGLE_ADS_ENABLE_MUTATIONS=true:
| Tool | Description |
|---|---|
update_campaign_status |
Pause/enable a campaign |
update_ad_group_status |
Pause/enable an ad group |
update_ad_status |
Pause/enable an ad |
update_campaign_budget |
Change daily budget amount |
add_negative_keywords |
Add negative keywords to a campaign |
Configuration
| Variable | Required | Default | Description |
|---|---|---|---|
GOOGLE_ADS_DEVELOPER_TOKEN |
Yes | — | API developer token |
GOOGLE_ADS_CREDENTIALS_PATH |
Yes | — | Path to OAuth client JSON |
GOOGLE_ADS_AUTH_TYPE |
No | oauth |
oauth or service_account |
GOOGLE_ADS_CUSTOMER_ID |
No | — | Default customer ID (skips passing it per tool) |
GOOGLE_ADS_LOGIN_CUSTOMER_ID |
No | — | MCC manager account ID |
GOOGLE_ADS_IMPERSONATION_EMAIL |
No | — | Service account impersonation email |
GOOGLE_ADS_ENABLE_MUTATIONS |
No | false |
Enable write tools |
GOOGLE_ADS_ENABLE_EXTENDED_TOOLS |
No | false |
Enable extended read tools (11 extra) |
GOOGLE_ADS_ENV_FILE |
No | .env |
Path to .env file (loaded if present, never overrides existing env) |
GOOGLE_ADS_API_VERSION |
No | v23 |
Google Ads API version |
Updates
Using npx @latest (recommended): You always get the latest version — no manual updates needed.
Using a binary: The server checks for new releases on startup and logs to stderr if outdated:
[mcp-gads] v0.2.0 available (current: v0.1.0). Download: https://github.com/pijusz/mcp-gads/releases/latest
Check your installed version:
mcp-gads --version
To update, download the new binary and replace the old one.
Development
Requires Bun.
git clone https://github.com/pijusz/mcp-gads.git
cd mcp-gads
bun install
bun test # tests
bun run build # standalone binary
bun run inspect # MCP Inspector
bun run check # biome format + lint
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.