AI assistants

MCP server

Rank.ai’s MCP server connects Claude Code, Codex, Cursor, Claude.ai and ChatGPT to your Rank.ai workspace. Add its address to your assistant, sign in, and ask about your AI visibility in plain words.

Server address: https://api.rank.ai/mcp
Sign in with: Your Rank.ai account, or an API key
Included on: The Growth and Agency plans and every done-for-you plan, including the Growth trial. See pricing

Setup

Connect your assistant

Pick your assistant. The first time it connects, it opens Rank.ai in your browser: sign in, pick the workspace and choose whether it may make changes. One connection works in one workspace.

One command, then sign in from Claude Code.

1
Add the server
Run this in your terminal. It adds Rank.ai to the project you are in; put --scope user after rank-ai to add it to every project.
Terminal
```
claude mcp add --transport http rank-ai https://api.rank.ai/mcp
```
2
Sign in
Start Claude Code, type /mcp, choose rank-ai and sign in to Rank.ai in your browser. Pick the workspace and whether it may make changes.
In Claude Code
```
/mcp
```
Ask
Ask in plain words, like “How often do ChatGPT, Claude and Gemini name us?”

Or connect with an API key

Keep the key in an environment variable and add the server to a project's .mcp.json:

.mcp.json

{
  "mcpServers": {
    "rank-ai": {
      "type": "http",
      "url": "https://api.rank.ai/mcp",
      "headers": { "Authorization": "Bearer ${RANK_AI_API_KEY}" }
    }
  }
}

Claude Code's MCP guide

Add it from the command line or in config.toml, then sign in.

1
Add the server
Run this in your terminal.
Terminal
```
codex mcp add rank-ai --url https://api.rank.ai/mcp
```
2
Sign in
Run the login command. Codex opens Rank.ai in your browser so you can sign in.
Terminal
```
codex mcp login rank-ai
```
3
Check it
In Codex, type /mcp to see rank-ai among your connected servers.
In Codex
```
/mcp
```

Or add it in config.toml

Codex reads its servers from ~/.codex/config.toml, or from .codex/config.toml in a project.

~/.codex/config.toml

[mcp_servers.rank-ai]
url = "https://api.rank.ai/mcp"

Or connect with an API key

Tell Codex which environment variable holds the key, then set it in the shell you start Codex from:

~/.codex/config.toml

[mcp_servers.rank-ai]
url = "https://api.rank.ai/mcp"
bearer_token_env_var = "RANK_AI_API_KEY"

Terminal

export RANK_AI_API_KEY="rnk_live_..."

Codex's MCP guide

Add it to mcp.json, then sign in.

1
Add the server
Open ~/.cursor/mcp.json to use Rank.ai in every project, or .cursor/mcp.json in one project, and add:
~/.cursor/mcp.json
```
{
  "mcpServers": {
    "rank-ai": {
      "url": "https://api.rank.ai/mcp"
    }
  }
}
```
Sign in
Save the file. Cursor opens the Rank.ai sign in the first time you use the server.
Ask
Ask in Agent. Cursor asks before it runs each tool, unless you tell it otherwise.

Or connect with an API key

Add the key as a header:

~/.cursor/mcp.json

{
  "mcpServers": {
    "rank-ai": {
      "url": "https://api.rank.ai/mcp",
      "headers": {
        "Authorization": "Bearer ${env:RANK_AI_API_KEY}"
      }
    }
  }
}

Cursor's MCP guide

A custom connector, on the web and in Claude Desktop.

1
Add the connector
In Claude, open Customize, then Connectors. Select +, then Add custom connector. Enter the name Rank.ai and the server address, then select Add.
Server address
```
https://api.rank.ai/mcp
```
Connect
Select Connect, sign in to Rank.ai, pick the workspace and select Allow.
Turn it on in a chat
In a chat, open the + menu, then Connectors, and switch on Rank.ai.

Anthropic's custom connector guide

A connection in developer mode.

Turn on developer mode
In ChatGPT, open Settings, then Security and login, and turn on Developer mode.
2
Add Rank.ai
Go to ChatGPT Plugins at chatgpt.com/plugins and select the plus button. Enter the name Rank.ai, a short description and the server address, with OAuth sign in.
MCP server URL
```
https://api.rank.ai/mcp
```
Sign in and ask
Create it, sign in to Rank.ai, pick the workspace and select Allow. Then start a new chat and add Rank.ai from the tools menu. ChatGPT asks you to confirm write actions.

OpenAI's guide to connecting ChatGPT

Any assistant that connects to MCP servers over HTTP.

1
Point it at the server
Use the address below with the HTTP transport (some clients call it streamable HTTP).
Server address
```
https://api.rank.ai/mcp
```
2
Sign in, or send a key
Clients that support MCP authorization find the sign in from the server's 401 response. Any other client sends an API key with every request:
Header
```
Authorization: Bearer rnk_live_...
```

The MCP specification

API keys

Use an API key

For your own scripts, or a client that can’t sign in, send an API key with every request.

In Rank.ai, open Settings, then API keys, and generate a key. Label it after the assistant or script that will use it, and keep it in an environment variable.
Copy the key when it appears. Rank.ai shows each key once.
Send it as a bearer token with every request:

Header

Authorization: Bearer rnk_live_...

Open API keys in Rank.ai

Examples

What you can ask

Ask the way you would ask a colleague, and your assistant picks the tools. These examples are for a med spa in Austin.

How often do ChatGPT, Claude and Gemini name us this month?
get_ai_visibility
Which questions are we losing, and which brands win them?
list_questionsget_question
Show me what ChatGPT said for the worst one, and the sources it used.
get_answer
Which sites does AI cite most for med spas in Austin?
list_cited_sources
Where do we rank on Google Maps for “med spa near me”?
get_map_scan
Draft an article to win “best Botox in Austin”.
draft_article
Grade prospect.com in AI answers.
grade_domainget_grade

Reference

Tools

Your assistant sees every tool when it connects and chooses the ones a question needs. These are the ones you will use most.

Your workspace

Tool	What it does	Access
get_account	Your plan and what it includes, credits, article allowance and what this connection may do.	Reads
get_overview	How you are doing across AI answers, Google, local and content, and what needs attention.	Reads

AI answers

Tool	What it does	Access
get_ai_visibility	The share of answers that name you and cite your site, per engine, against the brands named most.	Reads
list_questions	Your tracked questions, how often each engine names you, and which ones you are losing.	Reads
get_question	One question: each engine's latest answer, whether it names and cites you, and where.	Reads
list_answers	The answers to one question, newest first.	Reads
get_answer	One answer in full: its text, every brand and every source, and the searches the engine ran.	Reads
list_cited_sources	The sites AI cites for your questions, and their top pages.	Reads
suggest_questions	Questions worth tracking. Adds nothing.	Reads
add_questions	Track new questions.	Changes data
update_question	Reword, change engines, pause or resume a question.	Changes data
run_check	Ask the engines a question now.	Uses credits or allowance

Competitors

Tool	What it does	Access
list_competitors	Tracked competitors, brands AI names that you don't track, and who ranks in Google.	Reads
get_competitor	One competitor's Google, paid search and AI footprint.	Reads
track_competitor	Add a competitor to track.	Changes data
untrack_competitor	Stop tracking a competitor.	Changes data

Google and Maps

Tool	What it does	Access
list_keywords	Your Google positions, with the change and the ranking page.	Reads
get_keyword	One keyword's history and the results above you.	Reads
list_local_rankings	Map grid tracking for each location and keyword.	Reads
list_map_scans	Your saved map scans.	Reads
get_map_scan	One map scan: coverage for each keyword and the competitors on top.	Reads
list_audits	Your audits and monthly reports.	Reads
get_audit	One audit, a section at a time.	Reads

Content

Tool	What it does	Access
list_content_gaps	Keywords competitors win that you don't target.	Reads
list_articles	Your articles by status.	Reads
get_article	One article with its text.	Reads
draft_article	Write an article for a lost question, a content gap or a title.	Uses credits or allowance
publish_article	Publish an article to your connected site or hosted blog.	Changes data

AI grades

Tool	What it does	Access
grade_domain	Grade any domain in AI answers, a prospect's included.	Changes data
get_grade	The grade's result.	Reads

Limits

Credits and limits

Reading your data uses no credits. Changes and checks work the way they do in the app.

Changes wait for a yes. A tool that changes your workspace or uses credits first returns a preview: what will happen and what it costs. Nothing runs until your assistant calls it again to confirm, so it can ask you first.
Checks use your plan’s credits and drafts your article allowance, the same as in the app. Grading a domain uses no credits.
Each connection can make 60 calls a minute and 3,000 a day, and each workspace 120 a minute and 10,000 a day. When your assistant reaches a limit, the reply says when to try again.
Errors come back as a code and a sentence, such as [plan_required] or [rate_limited], so your assistant can tell you what to do next.

See how often ChatGPT, Claude and Gemini name you.

Enter your website. In about two minutes, Rank.ai asks ChatGPT, Claude and Gemini 12 questions your customers ask and grades how often they name you.

Your grade out of 100How often AI names you, cites your site, and how high it ranks you.
Who gets namedEvery competitor in the answers, most named first.
The pages AI readsThe sources behind each answer.
Three fixesWhat to fix first, with a brief for the first article.