CLI and MCP server
Your marketing engine. One tool call away.
RankControl wins Google rankings and AI citations for your site. The CLI and MCP server put every number and every action in your terminal, in Claude, in Cursor, wherever you already work. Setup is one command.
Agent session
MCP tools
0
one command to connect
Your AI assistant can read your code, draft your email, and plan your week. But your marketing platform is still a browser tab it cannot touch. Plug it in.
One command connects RankControl to Claude, Cursor, or any agent that speaks MCP, with approval gates that keep every publish in your hands. Built by a team doing SEO since 2009, including for multiple Fortune 1000 companies.
Ten tools, full loop
From visibility check to published article
npx rankcontrol mcp · 10 tools
get_visibility_score
Your overall AI visibility number, right now.
get_visibility_trend
Start hereDaily score history from your weekly citation checks.
list_citations
Recent checks per AI model, cited or not.
get_overview_funnel
AI crawls, AI visits, and AI leads, last 30 days.
get_planning_capacity
How many content slots are left this month.
plan_content
Propose article titles. Nothing gets scheduled.
commit_planned_titles
Approve reviewed titles onto the calendar.
list_content
Every article with its current status.
publish_content
Dry run by default. Confirm to go live.
list_jobs
What the agents are working on right now.
Included in the plan. Nothing extra to buy.
Start Free TrialHow it works
Your workspace, wherever you work
Use it as a plain command line tool, hand it to your AI assistant, or wire it into your own agents. The rules stay the same everywhere.
01 / Terminal
The dashboard, minus the dashboard.
Check your lead funnel and what every agent is working on without leaving the command line. Connect once, and every number you see in the dashboard answers from the same live workspace.
Terminal
$ npx rankcontrol funnel
ai crawls 412 last 30 days
ai visits 168
ai leads 9
$ npx rankcontrol jobs
forge writing: B2B Payment Tools
radar research sweep done ✓
Works in
Terminal, CI, anywhere
same limits as the dashboard
02 / MCP
Ask your assistant. Get your real numbers.
Add RankControl to Claude Code, Claude Desktop, Cursor, or Codex in about two minutes. Your assistant answers from your live data instead of guessing, and can plan next steps the moment it spots a gap.
Your AI assistant
How is our AI visibility this week?
get_visibility_trend …
Score is 64, up 6 this week. ChatGPT now cites you on 18 of 50 tracked queries. Want me to plan content for the 7 gaps?
Yes, plan them.
plan_content …
Real numbers
Straight from your workspace
no guessing, no hallucinating
03 / Guardrails
Agents propose. You approve.
Planning proposes titles, you commit the ones you want. Publishing is a dry run until you confirm. The gates live on our servers, so no client, script, or agent can route around them, and every action is logged.
Approval gates
server-sideEvery action logged. Quotas enforced on our servers.
Auto-published
Never
without your confirm
Works Where You Work
From Claude to Cursor to a plain terminal or your CI. One package covers every place you already work.
2 minto connect
Keys You Control
Log in with a browser approval, grant only the permissions you want, and revoke any key anytime. Read-only keys included.
100%of actions logged
Same Rules Everywhere
Quotas and approval gates run on our servers. The command line gets no shortcuts and no side doors.
0ways around approval
This morning
Auto-published
“Our agent checks citations every morning and drafts the next batch of titles. I review them over coffee and approve. It runs like an ops team that never sleeps.”
Ana Sørensen
Head of Growth, Driftline
0
tools your agents can call
0
command to connect any of them
0
publishes without your confirm
Common questions
Before you connect
Can an agent publish without me?
No. Publishing is a dry run by default, and nothing goes live until you explicitly confirm it. Planning is split the same way: agents propose titles, you approve them. Quotas and approval gates run on our servers, so no client can route around them.
Which AI tools does it work with?
Anything that speaks MCP: Claude Code, Claude Desktop, Cursor, and Codex, plus any custom agent. The CLI also works in a plain terminal or CI, no AI required.
Do I need to be a developer?
No. Connecting is one copy-paste command, and login is approved in your browser. And if you never touch a terminal, everything here is also in the dashboard.
Is my data safe?
Keys are scoped to exactly the permissions you grant, read-only keys are available, and you can revoke any key in Settings. Every action taken through a key is logged, so you always know who did what.
Does it cost extra?
No. The CLI and MCP server are included in the one RankControl plan, same data, same limits, same everything.
Free research report
The State of AI Visibility in 2026
25-page research document covering how brands get cited by ChatGPT, Perplexity, Claude, and Gemini — with data, benchmarks, and actionable strategies.
No spam. Unsubscribe anytime.
RankControl Research
The State of AI Visibility in 2026
ChatGPT · Perplexity · Claude · Gemini
Explore the rest of the platform
Give your AI tools the keys. Keep the approval gates.
One command connects your terminal, Claude, or Cursor. Free for your first 30 days.
