Fullrun — Google Ads For Ai Agents
LocalAbout Fullrun — Google Ads For Ai Agents
Fullrun is a CLI for AI agents and operators to inspect and optimize Google Ads accounts through the Fullrun API. It returns campaign diagnostics and metrics, and can trigger guarded optimization runs without storing Google Ads credentials locally.
Key features
- 01Prioritized account health reports
- 02Campaign and keyword listings
- 03Clicks, conversions, and CPA metrics
- 04AI-driven optimization runs
- 05Human-readable or JSON output
Use cases
- 01Find high-priority campaign problems
- 02Review keyword performance by campaign
- 03Run optimizations and verify the results
MCP support
Fullrun also provides an MCP server package, `@fullrun/mcp`, for MCP clients such as Claude Desktop and Cursor. The CLI itself communicates with the Fullrun API.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
npx -y @fullrun/mcpFULLRUN_API_KEYSecretYour Fullrun API key for authenticating with the Fullrun APILocal servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.
Tools
5Saved tool list · updated 3w ago
This tool takes no parameters.
Details
From the project README
Google Ads management CLI for AI agents.
Works with OpenClaw, Claude, and any agent that can run shell commands.
Fullrun is an AI agent that manages Google Ads for you. It monitors your campaigns around the clock, diagnoses problems, adjusts bids, adds missing ad extensions, pauses wasteful keywords, and creates new campaigns — all autonomously.
This CLI brings that same engine to your terminal. Instead of using the Fullrun dashboard, you (or your AI agent) can manage Google Ads with simple commands. Every command auto-detects your environment — colored output in a terminal, structured JSON when piped to another program.
npm install -g fullrun
fullrun login frun_YOUR_API_KEY
fullrun triage
What an agent can do
- Diagnose — Run
fullrun triageto get a prioritized health report (CRITICAL → HIGH → MEDIUM → LOW) - Inspect — List campaigns, keywords, and performance metrics
- Optimize — Trigger
fullrun runto let the AI agent fix the highest-priority issues - Repeat — Triage again to confirm the fixes worked
All the hard stuff — Google Ads API calls, bid calculations, budget guardrails, PPC best practices — happens server-side. The CLI is a thin client. No credentials stored on your machine.
Commands
| Command | Description |
|---|---|
fullrun login <key> | Authenticate with your API key |
fullrun triage | Account health report with prioritized issues |
fullrun campaigns:list | List campaigns with status, budget, and metrics |
fullrun performance | Account metrics — clicks, conversions, CPA |
fullrun keywords:list | Keywords with performance data |
fullrun run | Trigger a full AI-powered optimization run |
Options
--format json— Force JSON output (default when piped)--format human— Force colored terminal output (default in TTY)--days <n>— Look-back period for performance data (default 7, max 90)--campaign <id>— Filter keywords by campaign
Agent integration
OpenClaw
Install the CLI globally and OpenClaw auto-discovers it via the bundled SKILL.md:
npm install -g fullrun
Then tell your agent: "Check my Google Ads and fix anything that's underperforming."
Claude Code
The SKILL.md works as a Claude Code skill. The agent reads it and knows how to use every command.
MCP (Claude Desktop / Cursor)
{
"mcpServers": {
"fullrun": {
"command": "npx",
"args": ["@fullrun/mcp"],
"env": {
"FULLRUN_API_KEY": "frun_..."
}
}
}
}
Any agent
Any agent that can execute shell commands can use Fullrun. Output is auto-detected — JSON when piped, human-readable in a terminal.
Getting an API key
- Sign up at fullrun.app
- Connect your Google Ads account
- Go to Settings > API Keys
- Create a key and copy it
How it works
CLI / MCP / Agent
│
▼
Fullrun API (api key auth)
│
▼
Triage engine → Google Ads API
Guardrails, bid caps, PPC rules
The CLI never touches Google Ads directly. It calls the Fullrun API, which runs the same triage and optimization engine that powers the web dashboard. All guardrails (mutation limits, bid caps, budget safety) are enforced server-side.
Development
git clone https://github.com/tuckerschreiber/fullrun-cli
cd fullrun-cli
npm install
npx tsx src/bin/fullrun.ts --help
License
MIT
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.