You approve
Nothing changes until you approve and apply the proposal.
The Google Ads agent turns your own spend and conversion data into daily budget proposals. You approve each one, you set the caps, and nothing in the account changes until you say so.
ProThe Google Ads agent ships on Pro. Free shows you the gap it found.
daily proposals · your spend and conversions · caps you set
The problem
Queries stop converting, budgets stay put, and nobody turns the evidence into a proposal you can review.
Spend drifts to queries that never convert.
Budget decisions arrive without the conversion evidence beside them.
Nobody reviews the account often enough to propose the next move.
The Google Ads agent
Four steps. Every run. Nothing ships without you.
Nothing changes until you approve and apply the proposal.
Plays well with others
What it costs
Free shows you the measured gap. Pro, at $249 a month, turns on the Google Ads agent workflow. Nothing ships without your approval.
$0/mo
Know exactly what Google and the LLMs say about you. Fix it yourself.
5 one-time credits
The one-time grant never expires · No team seats · 1 website
Start freemeasures only
$129/mo
Eight agents draft the fixes and the posts for you. You approve.
2,000 credits each month
Monthly plan credits do not roll over · 3 team seats · 1 website
Choose Litemeasures only
$249/mo
All ten agents, Google Ads and technical docs included.
2,000 credits each month
Monthly plan credits do not roll over · 10 team seats · 1 website
Choose Proships on this plan
Built for developers
Copy it, paste it into your agent. It installs ai-cmo, pulls your score and ships the fixes as a PR you review.
Install the ai-cmo MCP server (npx -y ai-cmo-mcp@0.1.0), score ai-cmo.dev, and fix everything in my review queue. This one-liner is the quick start. The full guided setup for ai-cmo.dev begins with https://ai-cmo.dev/llms.txt and https://ai-cmo.dev/openapi.json, resolves the workspace with aicmo_whoami({}), and ends at the install page https://ai-cmo.dev/api/v1/workspaces/$WORKSPACE/install. Read https://ai-cmo.dev/llms.txt and https://ai-cmo.dev/openapi.json. If npx returns 404, report the unavailable package and stop. Run aicmo_whoami({}); expect authorized workspace access. For each candidate, set WORKSPACE to its returned id and read authenticated GET https://ai-cmo.dev/api/v1/workspaces/$WORKSPACE/settings with Authorization: Bearer $AICMO_API_KEY (settings:read), or use verified settings already supplied by the client. Expect domains to contain ai-cmo.dev; stop if no matching workspace or authorized settings are available. Run aicmo_score and aicmo_recs_list with the matching workspace id. Follow each returned recommendation’s brief and verification steps, run the repository checks and expect them to pass. Open a pull request for review; do not publish, deploy or merge.Fine, ask
The other nine
You build the product.
We make sure people find it.