Run Google Ads end to end from your favorite AI chat
Porter connects Google Ads to Claude, ChatGPT and other AI clients through one MCP server. MCP is the Model Context Protocol, the open standard AI apps use to reach other tools. Ask for a report, a search-term cleanup or a new campaign, and Claude works from your live Google Ads accounts.
Campaign management
Campaigns: build Search campaigns, Performance Max asset groups, Demand Gen and App ads.
Ads: write responsive search ads and add assets to the right ad group.
Keywords: add keywords and negatives, pause the ones that burn spend, change bids.
Budgets: shift budget between campaigns, with the plan shown before anything changes.
Multi-account and cross-channel: work in every advertiser account your Google login reaches, client accounts under a manager account included. The Meta Ads MCP and the Microsoft Ads connector for Claude run through the same server.
Audits: account structure, naming, budgets and tracking, ranked by what to fix first.
Alerts and schedules: a message when Cost / conv. passes your cap, and a weekly report by email.
Google Ads next to GA4: Cost and Clicks beside GA4 sessions and key events, in one table.
Keyword and search-term research
Search terms that waste spend: every Search campaign term with Clicks and Cost but no conversions.
Keyword ideas: Google’s keyword ideas for a seed term, grouped the way you plan ad groups.
Geo search: find location targets by name before you build a campaign.
Match type and query mining: compare each Search keyword match type with the Search term that triggered the ad.
Creative generation
Responsive search ad copy: headlines and descriptions written from your landing page and best performers.
Demand Gen images: image assets generated from a prompt or a product photo.
Video: a short video asset from a prompt or an image you already have.
Straight into the ad group: generated assets upload to Google Ads as part of the paused draft.
Reference
Google Ads MCP server reference
One MCP server: Google Ads fields, metrics, dimensions and actions reach Claude, ChatGPT and other AI clients through Porter’s MCP. The same server URL works in any MCP client that supports remote servers. The data comes from Porter’s Google Ads connector, which reads Google Ads through the official API.
Visibility: Cost · Impressions · Search Impr. share · Search Lost IS (budget) · Search Lost IS (rank)
Engagement: Clicks · CTR · Avg. CPC
Conversion: Total conversions · Cost / conv. · Conversions value · Conv. value / cost
Context: vs previous period · vs target · trend · budget pacing
Google Ads dimensions and segments
Campaign: Account name / description · Campaign · Advertising channel type · Campaign status
Keyword and search term: Search keyword (Criterion) · Search keyword match type · Search term · Keyword quality score
Audience and geo: Country/Territory · Device
Time: Date · Week · Month · Day of Week · Hour of day
Other destinations
The same Google Ads connection also sends data to Data Studio (formerly Looker Studio), BigQuery and Google Sheets, at no extra cost.
Google caps API traffic through Google Ads API access levels. Porter makes the calls through its own approved access, so you never apply for a level yourself.
Security and trust
Safety and security
Reports from Porter’s Google Ads MCP are read-only, and every change stays in your hands. New campaigns, ad groups and ads are created paused, and Claude or ChatGPT asks for your approval before Porter writes anything to your Google Ads account.
Read-only first
Ask for reports, audits and research without writing to Google Ads. Analysis and campaign changes are separate jobs, so a report can’t edit a campaign.
Reporting priced per account
Reporting is billed per connected data source account. Asking one more question about the same account adds no cost.
Approved Google sign-in
Connect with your Google login through Porter. You skip the developer token, the Google Cloud project and the API keys.
Manager accounts stay scoped
Porter works in the advertiser accounts your login can reach. From a manager account, each write goes to the client account you name.
Draft mode by default
Campaigns, ad groups and ads start paused, and nothing serves until you approve it or turn it on in Google Ads. Removing a campaign, keyword, budget, ad group or asset group is permanent, so Porter asks first.
Budget guardrails
Porter shows the budget plan before it applies, and the change waits for your yes. A wrong budget costs real money, so it gets its own checkpoint.
Official API, no browser automation
Porter works through the official Google Ads API. Nothing clicks around the Google Ads interface on your behalf.
Revoke access anytime
Disconnect Porter from your connection settings whenever you want. Access ends with the Google connection you approved.
One server URL connects Porter’s Google Ads MCP to every AI client: paste it once, sign in with Google and start asking. Claude installs it as a plugin or a custom connector. ChatGPT connects through the Porter Metrics app or developer mode, and both reach the same Google Ads accounts.
https://mcp.portermetrics.com/mcp
How to install Google Ads MCP in Claude
Use the Claude plugin for the guided route. Claude keeps the plugin synced when Porter ships new tools.
Open Claude Settings → Customize → Plugins.
Click Add → Add marketplace → Add from a repository.
Pasteportermetrics/porter.
Confirm Use portermetrics/porter.
Keep Sync automatically turned on.
Install the Porter Metrics card from the Directory.
Connect the Porter connector from the plugin settings and sign in with Google.
Ask Claude for Google Ads data, a search-term review or a campaign draft.
Sign in. Your browser opens: sign in with the Google account you use in Porter.
codex mcp login porter
Check:codex mcp list shows porter as logged in. Start codex. Ask: “Which Google Ads accounts can you see?”
The same URL works in Windsurf, Cline, Zed and any client that supports remote MCP servers over HTTP.
Official Google Ads MCP server
Google publishes its own server as code in the googleads/google-ads-mcp repository on GitHub. Google’s guide has you run it on your machine or deploy it to your own Cloud Run service, with your own Google Cloud project and OAuth client.
Use Google’s server for read-only GAQL access in a project you manage. Use Porter’s URL for writes with approval, ready-made fields and Google Ads next to your other channels. Both go into the same MCP Server URL field.
Try this first prompt once you’re connected:
Rank my Google Ads campaigns by Conv. value / cost for the last 7 full days and flag any whose Cost rose more than 20% versus the 7 days before.
Porter’s no-code walkthrough connects Google Ads to Claude through Porter and then asks Claude for campaign data. The steps match the Claude plugin and manual setup on this page, so you can follow along in text or on video.
Prompts
Google Ads MCP prompts
Porter’s Google Ads skills are ready-made MCP prompts that finish an operational task. Each one turns live account data into a report, a diagnosis, a research list, an alert or a paused draft you approve. Copy a prompt into Claude or ChatGPT, then change the dates or thresholds to fit your account.
Reporting skills
Connection test
List every Google Ads account I can access, then show Cost, Clicks, Total conversions and Cost / conv. by campaign for the last 7 days.
Wasted search terms
Find search terms from the last 30 days with Cost above $50 and zero Total conversions. Group them by campaign and suggest negatives, but don’t add them yet.
Impression share lost
For each Search campaign, show Search Impr. share, Search Lost IS (budget) and Search Lost IS (rank) for the last 30 days. Tell me which campaigns are held back by budget and which by rank.
Performance Max vs Search
Compare Performance Max and Search campaigns by Cost, Conversions value and Conv. value / cost for last month, split by Advertising channel type.
Quality Score check
List active keywords with Keyword quality score below 5 and Cost / conv. above the account average. For each, say whether to fix the ad, the landing page or the keyword.
Monthly pacing
Show month-to-date Cost by campaign against each budget and project where spend lands at month end.
Client account comparison
Compare every client account by Cost, Total conversions and Cost / conv. for the last 30 days, worst first.
Weekly report
Write a weekly Google Ads report in simple words: what changed, what worked, what got worse and what to do next.
Research skills
Keyword ideas
Get keyword ideas for ‘trail running shoes’ in the United States and group them by intent and ad group.
Spend to sessions and key events
Put Google Ads Cost and Clicks by campaign next to GA4 sessions and key events for the same campaigns, last 30 days.
Campaign QA before launch
Audit the paused campaign ‘Spring Sale Search’ before I turn it on. Check budget, bidding, locations, keywords, negatives, ad copy and final URLs.
From the wasted search terms list, add the ones I approve as exact-match negatives at campaign level.
Budget optimizer
Compare campaigns by Cost, Conv. value / cost and Search Lost IS (budget). Propose budget moves, and don’t apply anything until I approve.
Pause or rebid keywords
Find keywords with Cost above $100 and no conversions in the last 30 days. Propose pauses or lower bids, and wait for my approval.
Campaign from a brief
Build a Search campaign from the brief I paste next. Create the budget, campaign, ad groups, keywords and responsive search ads, all paused.
Alert agent
Every Monday, check active campaigns for Cost / conv. spikes, falling Conv. value / cost and spend with no conversions. Send me an alert when something needs review.
Who this is for
Who this is for
Porter’s Google Ads MCP is for marketers who run Google Ads and act on what they find. PPC managers, agencies, in-house teams and analysts use it to answer an account question in one prompt. The answer then becomes a change, a dashboard or an alert.
For marketers who report and then act.
Agencies and freelancers
Every client account: one question across the accounts under your manager account, with no logging in and out.
Client reports: a live report the client opens on Monday that you never rebuilt.
Account audits: the quarterly review opens with what to fix.
Bulk changes: one instruction updates keywords or budgets across accounts, each change approved.
In-house marketers
Cross-channel reporting: Google Ads next to your other ad platforms and the rest of Porter’s 30+ data sources.
Executive reports: one link that’s already current when leadership opens it.
Attribution: Google Ads Cost beside GA4 key events and the revenue in your CRM.
Budget pacing: month-end spend projected while there’s still time to move it.
PPC managers
Search terms: the wasted-terms list is ready before Monday’s check-in.
Impression share: you know whether budget or rank is holding a campaign back.
Keywords: pause, rebid or add negatives from the chat, approved one by one.
New campaigns: a brief turns into a paused campaign you review in Google Ads.
Analysts
Field-level access: the same field labels in chat, Sheets and BigQuery, so numbers match across tools.
Questions in words: ask for data in a sentence, with no GAQL to write or debug.
Warehouse destinations: send the same Google Ads data to BigQuery or Google Sheets.
Alerts: a spike reaches you the day it happens.
When to use Google’s own instead
You’re a developer who wants read-only GAQL access and is happy to run a Google Cloud project. Google’s server is first-party, open code and enough for that job.
Comparison
Official Google Ads MCP vs. Porter
Google’s official Google Ads MCP server and Porter’s both connect Google Ads to AI clients through MCP. Google’s Google Ads MCP server guide describes a read-only release that runs GAQL queries and that you host yourself. Porter’s server is hosted, writes with approval and adds your other channels.
Google’s server gives your AI read access to one platform. Porter gives it the paid media job: every channel, the reporting, the creative and skills that already know the work.
Both read live Google Ads data through Google’s official API, with OAuth sign-in. Here is where they part.
Porter
Google’s official MCP
Same job, different reach
Setup
Google sign-in and one server URL
Your Google Cloud project with API access, an OAuth client, then a local install or your own Cloud Run deploy
Developer token
Not needed
Not needed since Google’s sunset
Query style
Ready-made fields with the labels you see in Google Ads
GAQL queries
Manager accounts
Client accounts under your manager account, with each write sent to the one you name
The accounts your credentials can access
Writes
Campaigns, budgets, ad groups, keywords, ads and assets
Read-only in the current release
Approval controls
Created paused, approval before every write
No writes to approve
Price
Included in every Porter plan, the free plan too
Open code you host yourself
Only one side
Other channels
30+ data sources: Meta Ads, LinkedIn Ads, GA4, HubSpot, Shopify and more
—
Cross-channel questions in one conversation
Yes
—
Blending sources into one number
Yes
—
Live dashboards
Yes
—
Scheduled reports and alerts
Yes
—
Ad copy, image and video generation
Yes
—
Attribution to GA4 and CRM data
Yes
—
Data Studio, Google Sheets, BigQuery, Power BI
Yes
—
Skills for whole tasks
Yes
—
Google’s column follows the tools and modes in Google’s MCP server guide (as of November 2026).
Pick Google’s server for read-only GAQL work in your own Cloud project. Pick Porter when the job includes changes, other channels or reports someone else opens.
Access moved to the Cloud project: API access levels now sit on your Google Cloud project.
Older guides are out of date: tutorials that send you to the API Center for a token predate the change.
Google’s server still only reads: the current release runs GAQL queries and makes no changes.
Porter needs a Google sign-in only: no token, no Cloud project, no access-level application.
And the third-party MCPs?
Most other Google Ads MCP servers fall into three types, and the job tells you which one you’re looking at.
Reporting connectors: read your data into the chat and write nothing back.
Campaign controllers: create and edit, usually on one or two platforms.
Open-source servers: free to download, and yours to host, patch and update when Google changes the API.
Porter does all three jobs from one hosted server, starting on the free plan. Browse all marketing MCP servers Porter runs.
Google Ads MCP pricing
The Google Ads MCP is included in every Porter plan, the free-forever plan too. Reporting is billed per connected account.
14-day trial: everything unlimited, no credit card.
Free forever: unlimited connected accounts, 30 days of history (15 days in AI chats), 100 automation credits a month.
Paid Reporting: from $15 per account a month, or $12.50 billed annually.
Unlimited questions: AI chats, queries and dashboards are included. Reads are free; each completed change uses 1 automation credit.
Monthly billing
$40
Annual billing, per month
$33.33
Per account
$8.00
Reporting price for connected accounts, in USD. AI chats, queries and dashboards are unlimited. Annual billing is 10 months for 12.
The more accounts you connect, the lower the price per account. See the full calculator on Porter pricing.
Frequently asked questions
Is there an MCP for Google Ads?
Yes, more than one. Google publishes a read-only server you host yourself, and Porter runs a hosted one at https://mcp.portermetrics.com/mcp. Porter's reads Google Ads and makes changes you approve. The Porter MCP is included in every Porter plan, the free-forever plan too.
Does Google have an official Google Ads MCP?
Yes. Google's Google Ads MCP server is open code that runs GAQL queries and, in its current release, only reads data. You host it locally or on Cloud Run with your own Google Cloud project and OAuth client. Porter adds hosted setup, approved writes and your other channels.
Do I still need a Google Ads developer token?
No. Google sunset developer tokens on September 9, 2026, and Google Ads API access now comes from your Google Cloud project's access level. With Porter you skip both steps: sign in with Google, paste the server URL and ask your first question about your Google Ads accounts.
What is the Google Ads MCP server URL?
Porter's Google Ads MCP server URL is https://mcp.portermetrics.com/mcp. Paste it into Claude, ChatGPT or any MCP client as a custom connector, then sign in with Google. Google's own server is installed from its GitHub repository, so you use the address of your own local install or Cloud Run deploy.
How do I connect Claude to Google Ads?
Install the Porter plugin from Claude Settings → Customize → Plugins. Add the portermetrics/porter marketplace, install Porter Metrics, connect it and sign in with Google. You can also add a custom connector with Porter's server URL, which works on every Claude plan, Free included.
What is MCP in Google Ads? Is MCP an agent?
MCP is the Model Context Protocol, an open-source standard for connecting AI applications to external systems such as Google Ads. MCP isn't the agent. Claude or ChatGPT is the agent, and MCP is how it reaches your Google Ads accounts through Porter.
Is the Google Ads API free?
With Porter, the Google Ads API side is handled for you, so you never apply for API access or set up a Cloud project. You pay Porter per connected account for Reporting and by credits for Automation. The free-forever plan covers unlimited connected accounts and 100 credits a month.
Start free
Connect Google Ads to Claude or ChatGPT and run your first report, search-term review or paused campaign draft.
Start free: unlimited Google Ads accounts on the free-forever plan, or a 14-day trial with everything unlimited and no card.
