Skip to main content
MCP server CX Codex

citation-intelligence-mcp — Claude Code MCP server: what it does

citation-intelligence-mcp tells your agent which sources AI search engines cite for a query, and scores a page's odds of getting cited.

no license 229 installs updated 2026-05-26

Install into Claude Code · Platform CLI: remote server All methods and platforms ↓
claude mcp add --transport http citation-intelligence-mcp https://citation-intelligence-mcp--automatelab.run.tools
Open on GitHub

What citation-intelligence-mcp does

Point it at a query and it reports the URLs that Perplexity, ChatGPT, Claude, Gemini or Bing used as ground truth. domain.am_i_cited takes a batch of queries for one domain and returns per-query rank plus a citation rate. The Google side runs through signals.ai_overview, which needs SERPAPI_KEY.

Several tools never call an engine. citations.predict grades a URL from public signals (Wikipedia references, schema markup, llms.txt, canonical URLs, HTTPS) and returns ranked fixes. The audit.* group validates JSON-LD, builds an llms.txt from a sitemap, and probes whether GPTBot, ClaudeBot, PerplexityBot and other crawlers can actually fetch a page.

Answers land in a local cache, so domain.cited_for and domain.cited_for_diff replay stored citations and cost no API budget. Panels keep named query sets under <config>/panels/; panel.run writes a timestamped snapshot and citations.trend compares snapshots over time.

Call citations.check with no engine key and it fails, since it needs at least one. The predict score stays heuristic whatever else you configure.

What it helps with

  • see which sources an engine grounds its answer on
  • measure a domain's citation rate across a query set
  • match your page against top-cited competitors on the same signals
  • confirm AI crawlers are not blocked by robots.txt
  • generate an llms.txt from an existing sitemap

Example prompts for the agent

  • Run citations.provenance on the query “mcp server” and show which engines agree on the same URLs.
  • Score <url> with citations.predict and list the fixes it ranks highest.
  • Save a panel called editorial-watchlist with these queries, then run it and show the citation rate.

How to install citation-intelligence-mcp

Platform CLI: remote server

claude mcp add --transport http citation-intelligence-mcp https://citation-intelligence-mcp--automatelab.run.tools

Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more

Questions about citation-intelligence-mcp

How do I install citation-intelligence-mcp in Claude Code?

Run claude mcp add --transport http citation-intelligence-mcp https://citation-intelligence-mcp--automatelab.run.tools. That registers the hosted HTTP endpoint with Claude Code.

What does citation-intelligence-mcp do?

It collects the URLs AI engines cite for a query, tracks how often a domain shows up in those citations, and rates a page's citation likelihood from public signals. Other tools audit schema markup, sitemaps and crawler access.

Can I use it without API keys?

citations.predict, signals.wikipedia and the audit.* tools need no engine key. Citation checks want a key for at least one engine, signals.ai_overview wants SERPAPI_KEY, and signals.gsc_gap wants GCP service account credentials.

Similar MCP servers

All MCP servers for SEO and GEO →

Project sponsors

Companies that keep this analytics open