Citehound for Claude: connect the free MCP server

Connect it

Add the server address to Claude.

No key, no account. This is the address of the Citehound MCP server.

Server address

Claude on the web and desktop

Open Customize, then Connectors.
Click + Add, then Add custom connector.
Enter the name Citehound and paste the server address from above. Click Continue.
If Claude shows an authentication setting, choose No sign in. Citehound needs none. Then click Add.
In a chat, click the + button at the lower left, open Connectors and switch Citehound on for that conversation.

On Team and Enterprise plans an owner adds the connector first, under Organization settings, Connectors, Add, Custom, Web. Members then open Customize, Connectors and click Connect.

Claude Code

Run this once in a terminal, then check the connection with claude mcp list, or type /mcp inside a session.

Terminal

claude mcp add --transport http citehound https://getcitehound.com/api/mcp

Cursor

Add this to .cursor/mcp.json in a project, or to ~/.cursor/mcp.json to use it in every project.

.cursor/mcp.json

{
  "mcpServers": {
    "citehound": {
      "url": "https://getcitehound.com/api/mcp"
    }
  }
}

VS Code

Open the Command Palette and run MCP: Add Server.
Choose HTTP (HTTP or Server-Sent Events).
Paste the server address from above, then enter a server ID, for example citehound.

What you can ask

Eight things to type.

“Scan example.com and tell me what to fix first.”
scan_site
“Compare mysite.com with a competitor’s site. Where do they differ?”
compare_sites
“Write an llms.txt for my business. We make project software for small design studios.”
generate_llms_txt
“Write a robots.txt that allows ClaudeBot and GPTBot and blocks Bytespider, and name my sitemap.”
generate_robots_txt
“Give me Organization JSON-LD for example.com.”
generate_schema
“Show me the crawler access study and tell me what stands out.”
get_research
“What is the average AI readiness score for CRM software, and which checks do most sites fail?”
get_benchmark
“Which AI crawlers does Citehound track, and what does blocking each one mean?”
list_ai_crawlers

Tools

The 13 tools.

Every tool is read-only: it reads, calculates or writes text for you to use. None of them changes a site.

The 13 tools of the Citehound MCP server
Tool	What it does	Access
`scan_site`	Score any site out of 100 for AI readiness, with every failed check and its fix.	Read-only
`compare_sites`	Put two sites side by side and see every check where they differ.	Read-only
`generate_schema`	Get paste-ready JSON-LD for your site, with brackets where only you know the facts.	Read-only
`generate_robots_txt`	Allow or block named AI crawlers with a robots.txt you can paste as it is.	Read-only
`generate_llms_txt`	Draft an llms.txt, an optional file that describes your site and its key pages.	Read-only
`get_playbook`	Read the playbook for your vertical: the shift, three strategies and the pitfalls to avoid.	Read-only
`get_benchmark`	See where your score sits against the sites in your category.	Read-only
`list_ai_crawlers`	See the ten AI crawlers we track and what each one is for.	Read-only
`get_citation_prompts`	Get the buying-intent questions for your category to try in your own assistant.	Read-only
`get_citation_sample`	See a real, anonymized sample of a citation check, with the spread shown.	Read-only
`get_methodology`	See how the 100 points are built, check by check, and what the score cannot show.	Read-only
`get_research`	Read our published research, with the figures and how we got them.	Read-only
`get_sample_report`	Open a full-site report on our own site, before and after we fixed what it found.	Read-only

Limits

What it does not do.

The connector measures readiness: whether crawlers can reach a site and whether its pages carry signals a model can read. It does not measure whether any assistant names a brand. It has no Pro or signed-in layer yet.

Questions

Before you connect.

Is it free?

Yes. All 13 tools are free, with no account and no key. The tools that scan a live site are limited per domain, so one site is not fetched over and over. Claude’s own plan decides how many custom connectors you can add: the Free plan allows one.

Does it store anything?

Nothing about who you are. The server keeps short-lived counters keyed by a one-way hash of the scanned domain, which expire within an hour, to limit repeated fetches of one site.

What data does it send?

Claude sends Citehound the arguments of each tool call, such as a domain or a category name. The three tools that scan a live site (scan_site, compare_sites and generate_schema) fetch that domain’s public robots.txt, llms.txt, sitemap and homepage. The other 10 read Citehound’s own files and fetch nothing.

Which clients work?

Claude on the web and desktop, Claude Code, Cursor and VS Code are covered above. Another client that supports remote MCP servers over HTTP can use the same address. In Claude, Team and Enterprise owners add a custom connector for the organization first, then members connect to it.

Prefer to start in the browser?

The scan takes about 30 seconds and needs no email.

Run the free scan See the sample report

Questions, or a client that will not connect: hey@getcitehound.com

Ask Claude about your site’s AI readiness