How much does Google Ads MCP cost in 2026?
Google Ads MCP is completely free and open source. Zero subscription cost. But 'free' comes with hidden expenses: developer time for setup, a Google Cloud project, and ongoing self-hosting maintenance.
Last verified: May 2026
Pricing model
How Google Ads MCP pricing works
There is no pricing. Google Ads MCP is free, open-source software under the Apache 2.0 license. No subscription, no usage limits, no metering.
The actual costs are indirect:
Setup time: 15-60+ minutes of developer time for initial configuration. Requires Python 3.10+, a Google Cloud project, OAuth credentials, and a developer token.
Infrastructure (optional): If deploying to Cloud Run for remote access, you pay Google Cloud hosting fees (~$5-20/month for low usage). Local stdio mode is free.
Maintenance: Self-hosted means you're responsible for updates, security patches, and uptime. The project is v0.0.1 with 95 commits. Expect breaking changes.
AdKit charges $29-49/month and handles all of this for you: hosting, maintenance, updates, and support. Plus you get campaign management, competitor tracking, and creative tools that Google's MCP doesn't offer.
Plans
Plans and what each one includes
Google Ads MCP offers 1 tiers.
Open Source
$0
Apache 2.0 license
Free forever. Self-host the MCP server, query Google Ads data with GAQL.
- Google Ads data queries via GAQL
- Account listing and discovery
- Resource metadata for query building
- MCC/manager account support
- Read-only (no campaign creation or management)
- Google Ads only (no Meta, TikTok, LinkedIn)
- Self-hosting required (no managed option)
- Requires Python, Google Cloud project, developer token, OAuth setup
- No commercial support or SLA
- 3 tools only
Glossary
What these features actually mean
GAQL
Google Ads Query Language. A SQL-like language for querying Google Ads data. You need to know field names, resource types, and query structure. Example: SELECT campaign.name, metrics.clicks FROM campaign WHERE metrics.impressions > 1000. The MCP server's 'search' tool accepts raw GAQL, so your LLM must construct valid queries.
Developer token
A 22-character string from Google that grants API access. Requires a Google Ads Manager account and approval process. Explorer-level (minimum tier) is sufficient for the MCP server.
Self-hosted
You run the server on your own machine or cloud infrastructure. Google doesn't host it for you. Local mode uses stdio transport (fastest). Remote requires Cloud Run or similar with HTTP transport. You handle uptime, updates, and security.
Keep in mind
Hidden costs and fine print
Free in dollars, not in time
Setup takes 15-60+ minutes for developers. Cloud Run deployment can take days. Non-technical users effectively cannot use it. You're trading money for developer time.
Read-only limits what you can do
The canonical server cannot create campaigns, adjust bids, pause ads, or upload creative. The moment you want to take action on insights, you need a different tool.
Google Ads only
No Meta, TikTok, or LinkedIn support. Multi-platform advertisers need additional MCP servers, each with their own auth and maintenance.
3 tools is a narrow interface
The entire server exposes 3 tools. All the work of constructing valid GAQL queries falls on your LLM. If it gets the query wrong, you get cryptic API errors.
Explicitly experimental
Google states this is 'not intended for production use.' No SLA, no backward compatibility guarantees. It could change or be deprecated.
Real costs
What you'll actually pay
Google Ads MCP is free software. Hidden costs are developer time and optional infrastructure.
| Scenario | Plan | Monthly cost | Notes |
|---|---|---|---|
| Developer querying data locally | Open Source | $0/mo | Free. 15-20 min setup if experienced. |
| Remote access via Cloud Run | Open Source | ~$5-20/mo (Cloud Run) | Plus 30-60+ min deployment setup |
| Full ads workflow (create + manage) | Not possible | N/A | Read-only server. Needs separate tool for execution. |
Verdict
Is Google Ads MCP worth using?
If you're a developer who needs Google Ads data in your LLM workflows, yes. It's free, first-party, and gives you raw access to everything the Google Ads API exposes. No vendor lock-in, no subscription.
The question is what you do after reading the data. The MCP can tell you which campaigns underperform. It cannot pause them, adjust bids, or create new ones. For that, you need either manual work in Google Ads Manager or a different tool.
AdKit starts at $29/month for 1 account, MCP included. Read performance, create campaigns, manage bids, track competitors, browse 500,000+ ads, generate creative. Cross-platform (Meta, Google, TikTok). Two-minute setup. If reading Google data is the whole job, Google's free MCP covers it. If you want the full workflow, AdKit is built for that.
Consider AdKit
AdKit goes beyond data queries. Competitor research, creative generation, campaign management, and a draft system that reviews changes before they go live. From $29/month for 1 account, $49/month for unlimited.
See how AdKit comparesMeet the Google Ads MCP alternative that does more than read data
| Feature | | |
|---|---|---|
| Free Tier | 7-day free trial, full access | Completely free (open source) |
| Entry paid tier | $29/month (1 account, includes MCP) | $0 (no paid tier exists) |
| Hidden costs | None. Flat monthly pricing. | Google Cloud project costs (if using Cloud Run), developer time for setup and maintenance |
| MCP call limits | Unlimited | Unlimited (self-hosted) |
Pricing questions
Want the full picture?
Pricing is one piece. See how Google Ads MCP and AdKit compare on features, safety, and creative tools.