Through the Model Context Protocol (MCP), AI assistants use the tools and measured data of CiteBoost. You ask in plain language – the assistant calls the right tools, reads your stored results and sums them up for you: as a weekly report, client presentation or to-do list.
Server URL
https://app.citeboost.ai/api/mcp
Included in the Pro, Agency, GEO Pro and GEO Agency plans.
Connected in three steps
01
Add the URL
Add the server URL to your assistant as a connector or MCP server.
02
Sign in with CiteBoost
Claude opens the CiteBoost sign-in – you allow access with one click. Other clients use your API key.
03
Ask a question
Ask about visibility, sources or competitors. The assistant picks the tools by itself.
What you can ask
The assistant answers these questions with your real data from CiteBoost. The tool it calls is shown in brackets.
How visible was my brand in AI answers over the last 30 days – and which competitors get named more often? (ai_overview)
Which websites do AI models cite for my prompts where my brand is missing? (citation_sources)
What does GPTBot see on my homepage when it does not run JavaScript? (bot_view)
Setup
How to connect the common assistants:
Claude (claude.ai, desktop and mobile app)
Open Settings → Connectors → “Add custom connector”.
Enter the name “CiteBoost” and the server URL, leave the OAuth fields empty and save.
Click “Connect”, sign in with CiteBoost and allow access.
Turn on the CiteBoost connector in the chat and ask your first question.
Claude Code
Add it in the terminal with your API key:
claude mcp add --transport http citeboost https://app.citeboost.ai/api/mcp --header "x-api-key: <API-KEY>"
Cursor, VS Code and other MCP clients
Add this to your client’s MCP configuration (e.g. .cursor/mcp.json):
You create the API key in the app under “API access”. It is shown only once and stored only as a hash.
The 16 tools
The assistant decides from your question which tools to call – you do not need to know their names.
AI visibility
List prompts
list_geo_prompts
Shows your stored prompts – the questions CiteBoost regularly asks the AI models.
Run a prompt now
run_geo_prompt
Asks a stored prompt live to the AI models: is your brand named, in which position, with what tone and which sources?
Prompt history
prompt_history
All previous runs of a prompt with date, visibility, position, tone, competitors and cited domains.
Visibility over time
ai_overview
Average visibility, tone, position, mentions and trend – compared with your competitors and with the most cited sources.
Quick check
ai_visibility_check
Asks AI models for the best providers for a term and shows whether and where your brand appears.
Cited sources & gaps
citation_sources
Which websites AI models cite for your prompts – especially where your brand is missing.
AI crawlers & pages
Bot view
bot_view
What AI crawlers such as GPTBot, ClaudeBot and PerplexityBot see on a page without JavaScript – compared with the finished page, with robots.txt rules and an access test per bot.
Citability
citability_check
Checks whether a page is readable and citable for AI crawlers: content without JavaScript, robots.txt, structure, schema, FAQ, author and date.
Page audit
page_audit
Rates a page on structure, content and technical basics and lists the most important improvements by priority.
SEO
Keyword research
keyword_research
Keyword ideas with search intent, search volume, difficulty and a look at the current top results.
Rank check
rank_check
Which position your domain holds for a keyword in Google Search.
Competitor analysis
competitor_analysis
Compares two domains: facts from both homepages, strengths and gaps.
Content Explorer
content_explorer
Content ideas and example content for a topic from research with Google Search.
Keyword gap
keyword_gap
Keywords a competitor is found for and you are not – with recommendations for new content.
Backlink gap
backlink_gap
Websites that link to up to three competitors but not to you – with strength, spam score and prioritised outreach targets.
Internal links
link_analysis
Internal and external links of a page with anchor texts and an assessment of the linking.
Frequently asked questions
What is MCP?
The Model Context Protocol is an open standard that lets AI assistants use external tools and data. With the CiteBoost MCP server, your assistant uses your prompts, measured results and SEO tools right in the chat.
Which assistants can I connect?
Claude on claude.ai and in the desktop and mobile apps connects in one click with sign-in. Claude Code, Cursor, VS Code and other clients that support remote MCP servers over HTTP connect with your API key.
Do I need an API key?
Not for Claude – you sign in with CiteBoost and allow access. For other clients, create a key in the app under “API access” and add it as the “x-api-key” header.
What does a query via MCP cost?
Exactly the same as in the app: new measurements and analyses count against your plan quota or your credits. Reading stored results, histories and reports costs nothing extra.
Can the assistant delete or change anything?
No. It can run prompts and start analyses, but it cannot delete data or change settings.
How do I disconnect?
In the app under “API access → Connected apps” you disconnect any connection with one click. Access tokens expire after one hour and are renewed only for connected apps; API keys can be revoked at any time.
Your AI visibility, right in the chat
Create an account, set up a project, measure prompts – then query everything through your AI assistant.