# Mentionary (full guide for AI agents) > Mentionary measures how often AI assistants recommend a brand when buyers ask unbranded questions. It writes the questions buyers ask AI from a website's homepage, asks ChatGPT, Gemini and Claude each one, and reports the mention rate per engine, the competitors named instead, the sources cited and a prioritised list of fixes. Google AI Overviews is in pilot. Perplexity is next. ## Connect - Remote MCP server (Streamable HTTP): https://platypus.mentionary.ai/mcp - Authentication: OAuth 2.1 with dynamic client registration and PKCE. The person you work for signs in to Mentionary once and approves the connection. There is no API key. - Add to Claude: https://claude.ai/customize/connectors?modal=add-custom-connector&connectorName=Mentionary&connectorUrl=https%3A%2F%2Fplatypus.mentionary.ai%2Fmcp - Claude Code: `claude mcp add --transport http mentionary https://platypus.mentionary.ai/mcp` - ChatGPT (web): open https://chatgpt.com/plugins, add a custom MCP server with the URL above and choose OAuth. - Web app sign-up: https://app.mentionary.ai/auth/sign-up - A new account starts on the Free plan. No card is needed for a first full analysis. ## Notes for agents - A workspace tracks one website the user owns or manages. Call list_workspaces first for a returning user. - Analyses run in the background. Poll get_setup_status until the stage you need has finished, then read results. - Plan limits apply to active buyer questions, monthly answer checks and competitors. When a limit is reached, the account owner can upgrade in the web app. - Results cover ChatGPT, Gemini and Claude. Do not report other engines as measured. ## Tools - `list_workspaces` (read): List the websites this account tracks - `create_workspace` (write): Start tracking a website you own or manage - `get_setup_status` (read): Check setup and analysis progress - `get_business_profile` (read): Read the detected personas, offerings and competitors - `confirm_business_profile` (write): Save the value proposition and personas - `add_personas` (write): Add buyer personas in one call - `get_question_guidelines` (read): What your assistant needs to write the buyer questions - `add_queries` (write): Save the questions your assistant writes, checked for bias - `generate_queries` (write): Or let Mentionary write the buyer questions - `list_queries` (read): List tracked buyer questions - `set_queries_active` (write): Turn tracking on or off per question - `activate_recommended_queries` (write): Fill the plan with the best mix of questions - `start_ai_visibility_analysis` (write): Ask ChatGPT, Gemini and Claude every active question - `get_visibility_brief` (read): The headline, key findings and chart-ready numbers - `get_ai_visibility_results` (read): Mention rate per engine, rivals named, cited sources - `set_competitors` (write): Set competitors and target markets - `get_citation_sources` (read): Your cited pages and the third-party sources engines cite - `get_recommended_actions` (read): Prioritised fixes and content gaps - `complete_setup` (write): Finish workspace setup ## Playbooks Chained prompts for common jobs. Run the steps in order in one conversation. ### Write the page AI should be quoting For: Founders, growth and content teams. Before you write a word, see which competitors the engines name for your category and which sources they lean on. Then draft the page for the question you are losing. 1. Map who wins the answer. Prompt: "Use Mentionary to set up a workspace for yourdomain.com and run the AI visibility analysis. Then list the buyer questions where a competitor is named and we are not, with the sources ChatGPT, Gemini and Claude cite for each." Returns: Competitors named per question and engine, with the cited sources. 2. Draft the page. Prompt: "From those Mentionary results, take the question where we trail most and draft a landing page for it: a headline, three proof points, a comparison table against the competitors named and an FAQ that covers what the cited sources cover." Returns: A page draft aimed at the exact question you are losing. ### Keep a live site in the answer For: SEO leads and marketing teams. Re-check the same buyer questions, see what moved, and turn the gaps into work: fixes for your pages and the third-party sites worth earning a mention on. 1. See what moved. Prompt: "Use Mentionary to pull the latest AI visibility results for yourdomain.com and tell me what changed: mention rate per engine, any competitor that gained ground, and the one thing to do this week." Returns: Mention rate per engine, rivals gaining ground, one action for the week. 2. Find the sources to win. Prompt: "Use Mentionary to list the sites ChatGPT, Gemini and Claude cite in answers where a competitor is named and yourdomain.com is not. Rank them by how often they are cited and suggest how we could earn a mention on the top five." Returns: Cited sites ranked by pull, with a plan for the top five. 3. Ship the fixes. Prompt: "Use Mentionary to get the recommended fixes for yourdomain.com and rewrite each one as a ticket: a title, why it matters for AI visibility, acceptance criteria and a priority." Returns: Tickets with a reason, acceptance criteria and a priority. Automatic weekly re-analysis is included in Pro and Agency. ### A board-ready AI visibility report For: CMOs, founders and marketing leads. One page leadership can read before the meeting: where you stand on each engine, who is named instead, and the three fixes worth funding. Then the slides. 1. Write the one-pager. Prompt: "Use Mentionary to measure how often ChatGPT, Gemini and Claude recommend yourdomain.com when buyers ask about our category, then write a one-page report for my board: mention rate per engine, the competitors named instead of us, and the three fixes to make first." Returns: Mention rate per engine, competitors named instead, three fixes. 2. Turn it into slides. Prompt: "Turn that Mentionary report into five slides for the board: the headline number, mention rate per engine, the competitors named instead, the three fixes with an owner each, and what we will measure next quarter." Returns: A five-slide outline ready for your template. ### Walk into every client call with numbers For: Agencies and consultants. One workspace per client site. Run the analysis, get a brief with the rivals named and three talking points, and a status line across every account. 1. Brief one client. Prompt: "Use Mentionary to set up a workspace for clientdomain.com, a site I manage, run the AI visibility analysis and write a one-page brief for our call: mention rate per engine, the competitors named instead, and three talking points." Returns: Mention rate, rivals named and three talking points for the call. 2. Check every account. Prompt: "Use Mentionary to list the workspaces on this account and give me one line per client site: mention rate, the competitor named most often instead, and whether setup is complete." Returns: A status line per client site. The Agency plan covers 500 buyer questions and unlimited competitors. ## Use-case directory ### Competitor research - Who AI recommends instead of you: The competitors named in buyer answers, ranked by how often. Prompt: "Use Mentionary to show which competitors ChatGPT, Gemini and Claude name when buyers ask about our category instead of yourdomain.com, ranked by how often each is named." - Engine by engine: Where you are strong, where you are missing, and how the three engines differ. Prompt: "Use Mentionary to compare how ChatGPT, Gemini and Claude treat yourdomain.com: mention rate on each engine, the competitor each one favours, and what explains the gap." - Where a rival wins the answer: The buyer questions a competitor wins and you do not, with the sources cited. Prompt: "Use Mentionary to list the buyer questions where a competitor is named and yourdomain.com is not, grouped by competitor, with the sources the engines cite in those answers." ### Pages and content - A landing page for the question you lose: A page draft aimed at the question you trail on, written against the cited sources. Prompt: "Use Mentionary to find the buyer question where yourdomain.com is furthest behind its competitors, then draft a landing page that answers that question better than the sources the engines cite today." - An FAQ built from buyer questions: Answers to the exact questions buyers ask AI, ready for your page. Prompt: "Use Mentionary to list the buyer questions where yourdomain.com is not mentioned, then write an FAQ section for our site that answers the ten most important ones in plain language." - A comparison page against your top rival: An honest comparison page aimed at the rival the engines name most. Prompt: "Use Mentionary to find the competitor named most often instead of yourdomain.com, then draft a comparison page against them: who each is for, a feature table and the questions buyers ask when choosing." ### Sources and links - Sites worth earning a mention on: The third-party sites the engines cite where rivals appear and you do not. Prompt: "Use Mentionary to list the sites ChatGPT, Gemini and Claude cite in answers where a competitor is named and yourdomain.com is not. Rank them by how often they are cited and suggest how we could earn a mention on the top five." - Where AI gets its information about you: Which of your pages the engines cite, the third-party sites they cite most and how it moves week by week. Prompt: "Use Mentionary to show which of yourdomain.com's pages ChatGPT, Gemini and Claude cite, the third-party sites they cite most with their authority, and how citations moved week by week. Then suggest which pages to strengthen and which sites to target." - Content gaps against competitors: Topics your competitors cover that the engines draw on, and you have not written. Prompt: "Use Mentionary to get the content gaps and recommended actions for yourdomain.com, then group them into a content plan for the next month with one page per week." ### Reporting - A board-ready AI visibility report: One page: mention rate per engine, competitors named instead, three fixes to fund. Prompt: "Use Mentionary to measure how often ChatGPT, Gemini and Claude recommend yourdomain.com when buyers ask about our category, then write a one-page report for my board: mention rate per engine, the competitors named instead of us, and the three fixes to make first." - The Monday visibility check: What moved since the last run, who gained ground, one thing to do this week. Prompt: "Use Mentionary to pull the latest AI visibility results for yourdomain.com and tell me what changed: mention rate per engine, any competitor that gained ground, and the one thing to do this week." - Five slides for leadership: A slide outline from measured numbers, ready for your template. Prompt: "Use Mentionary to pull the AI visibility results for yourdomain.com and outline five slides for leadership: the headline number, mention rate per engine, the competitors named instead, three fixes with owners, and what we will measure next." ### Teams and agencies - Fixes written as tickets: Each fix as a ticket with a reason, acceptance criteria and a priority. Prompt: "Use Mentionary to get the recommended fixes for yourdomain.com and rewrite each one as a ticket: a title, why it matters for AI visibility, acceptance criteria and a priority." - A client brief before the call: Measured numbers for a site you manage, with three talking points. Prompt: "Use Mentionary to set up a workspace for clientdomain.com, a site I manage, run the AI visibility analysis and write a one-page brief for our call: mention rate per engine, the competitors named instead, and three talking points." ### Agents and developers - Set up tracking from one message: The assistant runs the whole setup and reports back when results are in. Prompt: "Use Mentionary to set up AI visibility tracking for yourdomain.com: create the workspace, confirm the business profile, generate buyer questions, keep the ten most relevant active, set our top competitors and run the analysis. Tell me when the results are ready." - Fix the page in your repo (Claude Code): Claude Code reads your landing page, checks Mentionary and edits the copy. Prompt: "Read our landing page in this repo. Then use Mentionary to find the buyer questions where yourdomain.com is not named and the sources the engines cite for them, and edit the page copy so it answers those questions directly." ## Pricing - Free: $0, no card. 1 website, 10 tracked buyer questions, 1 full analysis across the three engines, 2 competitors. - Starter: $29 a month. 25 tracked buyer questions, 150 answer checks a month, 2 competitors, Re-run the analysis when you choose. - Pro: $99 a month. 100 tracked buyer questions, 1,000 answer checks a month, 10 competitors, Automatic weekly re-analysis. - Agency: $299 a month. 500 tracked buyer questions, 5,000 answer checks a month, Unlimited competitors, Automatic weekly re-analysis. ## FAQ ### Can't I just ask ChatGPT about my brand? You can, and you will get one polite answer to a question no buyer asks. Buyers ask for the best option in a category without naming you. Mentionary writes those unbranded questions from your site, asks ChatGPT, Gemini and Claude each one, and counts how often you are named and who is named instead. One answer is an anecdote. A rate across every question and engine is a measurement. ### What does Mentionary do? It reads your homepage, works out who buys from you, writes the questions those buyers put to AI, and asks ChatGPT, Gemini and Claude every one. You get your mention rate per engine, the competitors named instead, the sources the engines cite and a prioritised list of fixes. ### How do I use it in Claude or ChatGPT? In Claude, use the Add to Claude button and sign in once. In ChatGPT on the web, open chatgpt.com/plugins, choose Add custom MCP server, paste https://platypus.mentionary.ai/mcp and pick OAuth. Then ask in plain language. There is no API key. The MCP page has the steps for Claude Code and other MCP clients too. ### Which AI engines do you measure? ChatGPT, Gemini and Claude are measured on every plan. Google AI Overviews is built and running in pilot on selected workspaces. Perplexity is next. ### What does it cost? The Free plan costs nothing and needs no card: 1 website, 10 tracked buyer questions, 1 full analysis across the three engines and 2 competitors. Paid plans are Starter at $29 a month, Pro at $99 and Agency at $299. Paid trials last 14 days and start at checkout. ### Can I analyse a competitor's website? A workspace tracks a site you own or manage. Competitors are added to that workspace, so you see how often each one is named in the answers to your buyers' questions. ## More - Home: https://mentionary.ai/ - MCP connector guide: https://mentionary.ai/mcp - Methodology: https://mentionary.ai/methodology - Pricing: https://mentionary.ai/pricing - Short guide: https://mentionary.ai/llms.txt