CitedOrNot MCP server
Ask your AI assistant whether ChatGPT, Gemini, Perplexity and other AI engines recommend your
brand, which competitors they name instead, and what to publish to get named.
CitedOrNot checks, on a schedule, whether AI engines
name your brand when buyers ask what to buy or who to use. This MCP server lets an assistant such
as Claude, Cursor or VS Code read those results directly.
This repository is documentation only. The server runs at https://citedornot.com/mcp and
is part of the CitedOrNot product; its source code is not published here. This repo exists so
you can see what it does and how to connect to it.
What you need
- A CitedOrNot account on the Pro or Agency plan
(plans and prices).
- A personal MCP token: in CitedOrNot go to Settings → MCP and create one. It starts with
mcp_.
New to CitedOrNot? The free check needs no signup:
citedornot.com.
What you can ask
- "Which of my brands is named least often by the AI engines this month?"
- "Show the questions where a competitor was named and we weren't, on Perplexity."
- "How do the engines differ in tone about ?"
- "What should I publish next to get named more often?"
- "Which of my agency clients has the lowest visibility?" (Agency plan)
| Tool | What it returns |
|---|
list-brands-tool | Every brand on the account, with question counts and when each was last scanned. Call this first. |
get-brand-visibility-tool | Visibility rate, recommendation rate, sentiment split, a breakdown by engine and by question, and the competitor gaps. |
list-mentions-tool | Individual engine answers: the engine, whether the brand was named, tone, how strongly it was recommended, an excerpt and any citations. Filter by question, engine, tone or gaps only. |
get-recommendations-tool | The current "what to publish" advice: which competitor is winning, the query, proof points and action steps. |
list-clients-tool | Agency clients with their brands and a rolled-up visibility rate. Appears only for accounts that have clients. |
Every tool is read-only. None of them starts a new scan, uses your scan allowance or changes
anything on your account. The live list is also published as a
server card.
Connect
The endpoint is https://citedornot.com/mcp (streamable HTTP). Send your token as
Authorization: Bearer <your-token>, or as X-API-Key: <your-token> if your client can't set an
Authorization header. There is no OAuth sign-in.
Claude Code
claude mcp add citedornot --transport http https://citedornot.com/mcp --header "Authorization: Bearer <your-token>"
Claude.ai / Claude Desktop
Settings → Connectors → Add custom connector, paste the URL above. When it asks about
authentication choose None (this server uses a token, not OAuth), then add a header:
- Name:
Authorization
- Value:
Bearer <your-token>
Cursor / VS Code / Windsurf
Add a remote MCP server with:
- URL:
https://citedornot.com/mcp
- Header:
Authorization: Bearer <your-token>
Smithery
A listing with a token field is on Smithery:
citedornot/ai-visibility.
Limits and access
- 60 requests per minute.
- The token works only while the account is on Pro or Agency. It stops working if the plan is
downgraded or you deactivate the token under Settings → MCP.
- Treat the token like a password. Anyone who has it can read your brands' results.
Questions
hello@citedornot.com · citedornot.com
Licence
The text in this repository is released under the MIT licence (see LICENSE). That covers this
documentation only; the CitedOrNot service and its name and logo are not covered by it.