---
title: "MCP server"
description: "Give Claude, Cursor, or any MCP client your RankControl workspace as tools."
---

> Documentation Index
> Fetch the complete documentation index at: https://rctrl.com/llms.txt
> Use this file to discover all available pages before exploring further.

# MCP server

The same package that powers the CLI is an MCP server. Any MCP client gets
70+ tools over your workspace: visibility data, the content pipeline,
repurposing, link building, brand memory, and reports, all under the scopes
of the key it runs with.

## Setup

Authorize once (`rankcontrol login`, see [Getting started](/docs/getting-started)),
then register the server.

### Claude Code

```sh
claude mcp add rankcontrol -- npx rankcontrol mcp
```
### Claude Desktop / Cursor

```json
{
  "mcpServers": {
    "rankcontrol": {
      "command": "npx",
      "args": ["rankcontrol", "mcp"]
    }
  }
}
```
### Explicit key

```json
{
  "mcpServers": {
    "rankcontrol": {
      "command": "npx",
      "args": ["rankcontrol", "mcp"],
      "env": { "RANKCONTROL_API_KEY": "rctrl_pk_..." }
    }
  }
}
```

## Prefer a skill?

Agents that run shell commands directly (Claude Code, OpenClaw, and any
[skills](https://skills.sh)-compatible agent) can learn the CLI itself instead
of connecting over MCP:

```sh
npx skills add rankcontrol/rankcontrol
```

The skill carries the full command reference plus the ground rules: dry-run
semantics, the review gates, and when not to add `--confirm`. Source at
[github.com/rankcontrol/rankcontrol](https://github.com/rankcontrol/rankcontrol).

## What agents can do

| Area | Tools |
| --- | --- |
| Visibility | score, trend, share of voice, cited sources, sentiment, citation checks, tracked queries, crawler access |
| Content | list, scored ideas, plan ideas and titles, commit, generate, publish, reschedule, archive, citability report, internal links |
| Reports | exec summary, top wins, agent activity, page engagement, traffic, leads, funnel |
| Repurposing | queue, generate drafts, edit, Postiz channels, push, mark posted |
| Link building | backlinks and stats, outreach pipeline, contact finding, reply drafting, queueing, Link Network |
| Social | thread prospects, stats, status, reply drafting |
| Brand memory | read everything, update voice and style, manage products and buyer profiles |
| Team | list, invite, revoke, remove |
| Support | `contact_support` files a report to the team when something breaks |

Example prompts that work end to end:

> "Which of my tracked queries cite competitors but not us? Plan articles for
> the top three."

> "Update my brand tone to be more direct, then show me what the writers will
> see."

> "Summarize last month vs the month before and draft a LinkedIn post about
> the biggest win."

## Safety model

Agents get the same guardrails as humans, enforced server-side:

- **Dry run by default.** Publicly visible actions (publish, social push,
  outreach send, Link Network changes, team invites) return a preview until
  called again with `confirm: true`. Tool descriptions tell the agent a human
  should approve first.
- **Scoped keys.** A key without `write:publish` cannot publish, no matter
  what the agent asks. Grant the minimum during `login --scopes`.
- **Rate limits.** Every endpoint carries a per-key cap generous enough for
  real work and tight enough that a looping agent cannot run away with your
  usage. LLM-spending tools sit at 5 to 10 calls per minute.
- **Revocable.** Every key is visible in Settings → API and dies the moment
  you revoke it.

> **Caution**
>
> Treat an MCP-connected agent like a team member with exactly the key's
> scopes. The dry-run gates stop accidents, not intent, so scope keys down
> for agents you do not fully control.

Source: https://rctrl.com/mcp/index.mdx