
Query your AI visibility, competitors, sentiment and recommendations from inside Claude and other AI assistants, in natural language, using the Martekio connector.
Martekio now exposes a connector (a Model Context Protocol, or MCP, server) that your AI assistant can talk to. Once connected, you can ask the assistant questions about your brand's AI visibility and it will fetch live answers from your Martekio space, then write them up however you need: a summary, a competitor briefing, a monthly client report.
The connector is read-only. Assistants can see your Martekio insights but can never change anything in your account. It can report on any space you have access to, and access is revocable at any time.
| Assistant | How it connects |
|---|---|
| Claude on the web (claude.ai) | Sign in: paste the URL, log in to Martekio, pick a space |
| ChatGPT (chatgpt.com) | Sign in: paste the URL, log in to Martekio, pick a space |
Both assistants use a sign-in flow: you paste one URL and approve access with your normal Martekio login, no token to copy.
In Martekio, open Help & Settings → Connect AI Assistants (under Connections). You will see:
No token needed: you paste one URL, then approve access with your Martekio login and choose which space to share.
https://api.martekio.com/mcpYou have now connected Claude to Martekio's visibility insights.
Note that a paid ChatGPT plan is required to add MCP connectors.
https://api.martekio.com/mcp as the MCP server, and add it.You never call the tools yourself. You just talk to the assistant in natural language and it decides which tools to call, often chaining several to answer one question. Some examples, grouped by what you are trying to do:
If you have access to more than one space, you can pit two of your brands against each other. The assistant pulls each brand's figures and lays them side by side.
Twenty-eight read-only tools. You do not need to remember these; they are listed so you know what the connector can reach. If you would rather ask than read, "what can you do with my Martekio data?" makes the assistant call get_capabilities and present the same thing as a menu.
| Tool | What it returns |
|---|---|
| list_my_spaces | The brands this connection can report on, and which is the default. Only relevant if you manage more than one. |
| get_space_profile | Which brand a report is about: name, websites, country, and the brand identifiers that make a prompt branded. |
| get_sov_summary | Latest Share of Voice per assistant, total, branded and non-branded. |
| get_sov_trends | Share of Voice over time per assistant, to describe movement. |
| get_top_competitors | Your Competitive Threats ranked by how often they are cited, with your own citations for comparison. (Flag your rivals as Competitive Threats in Curate Competitors, on the Competition Explorer, first.) |
| get_citation_gaps | Prompts your Competitive Threats are cited on while you are not cited at all. |
| get_chatgpt_ads | Sponsored ads ChatGPT showed for your tracked topics, by advertiser, with served vs shown counts and whether each is your own brand or a competitor. Ask "which advertisers are buying ChatGPT ads against my topics?" |
| get_sentiment | How favorably assistants speak about you, per assistant, with an average score. |
| get_prompt_group_performance | Your Share of Voice, visibility and sentiment for each Prompt Category or intent in your latest analysis, branded and non-branded. Ask "what is my share of voice for each prompt category?" or "which intents am I weakest on?" |
| get_industry_insights | How you perform against your industry peers (Industry Insights): your own all-time metrics next to the peer average for your sector, per assistant, with the gap between them; the percentage of non-branded and branded responses that cite or mention you, plus average citations and mentions per response. |
| get_google_ai_overviews | Google AI Overview performance: presence, visibility and citation rates (all/branded/non-branded), plus how your organic ranking converts into AI Overview citations (fully visible / ranking overlooked / AI-only / invisible). |
| get_recommendations | Current visibility recommendations, prioritized. During a free trial this returns the highest-ranked recommendations only, with a note explaining the rest unlock on subscription. |
| get_impact_actions | Your recommendations grouped by the work they describe, ranked by how many separate analyses reached the same conclusion. Ask for this when you want to know what to prioritize or what keeps coming back, rather than one analysis's list. During a free trial this returns the top-ranked actions only, with a note explaining the rest unlock on subscription. |
| get_backlog | Your Visibility Backlog: the work you have actually picked up, as opposed to what was suggested. Items come from recommendations, from weekly tips, or typed in by hand, each with whether it is done, when it was added and who completed it. Ask it what you are working on, what has been finished, or whether a recommendation was acted on. During a free trial the items that came from recommendations are not listed, for the same reason the recommendations themselves are previewed; your own manual items and weekly tips always are. |
| get_agentic_traffic | AI bot (crawler) visits to your website (Agentic Traffic). |
| get_ai_referrals | Human visitors arriving from an AI assistant: total, per source, top landing pages, and their share of your site's traffic. |
| get_influential_sources | Where AI assistants get their information: how many responses cite each of your own influential sources (the list you curate in Manage Influential Sources), with per-assistant citation counts and unique URLs. |
| get_top_subreddits | The subreddits AI assistants cite most for your topics, each split by branded and non-branded queries with the actual Reddit post URLs on each side. |
| get_reddit_threads | The Reddit threads AI assistants cite: subreddit, post title and date, an excerpt of the post, which assistants cited it (split branded vs non-branded), how many people took part and who started it. |
| get_cited_page_health | The technical health of the pages AI cites: your pages compared with competitors' pages on load time, broken links, meta tags and structured data, with any broken pages of yours flagged first. |
| get_ai_responses | The actual AI reply texts, with sentiment and citation flags. |
| get_analysis_sessions | Your recent analysis runs: date, keywords processed, assistants covered. |
| get_collaborators | The people with access to the space: name, email and role. |
| get_tracked_prompts | The keywords and prompts you track, optionally filtered branded or non-branded. (Named get_tracked_phrases until 28 September 2026; a saved prompt using the old name needs updating.) |
| get_prompt_movers | (Named get_phrase_movers until 28 September 2026.) Which of your tracked prompts are gaining or losing AI visibility: each one's citation rate now against the previous equal window, the change in percentage points, a trend label, and a per-assistant breakdown. Prompts with too few answers in either window to be reliable are left out. |
| get_visibility_report_prompt | The ready-to-run AI Visibility Report prompt for the space, so you can generate the full branded report in the assistant itself. Martekio offers several named reports: ask for one by name, or leave it out for the default. It returns your own saved version of that report when you have one, or the standard one when you do not. |
| get_brand_logo | Your brand logo as embeddable image bytes. Used automatically when a generated report needs your logo, because fetching it from a URL fails inside a sandboxed assistant. |
| get_capabilities | A catalog of everything this connector can do, grouped by area, each with an example question. Ask "what can you do with my Martekio data?" to get it as a menu. |
| get_documentation | How Martekio itself works: what a feature does, how to read a dashboard, what a metric means, how to improve your visibility. Ask "how does Share of Voice work?" or "what is the Competition Explorer?". This is product help, not your account numbers. |
Every tool is time-windowed where relevant (default: the last 30 days, or 90 for trends) and you can just ask for a different period, for example "over the last quarter".
If your Martekio account has access to more than one space (generally the case if you have an Agency or Group subscription containing multiple spaces), a single connection can report on all of them, you do not need a connection per brand. On the sign-in flow you pick a default space; after that you switch simply by naming the brand:
To see which brands you can query, just ask the assistant "which brands can this connector see?" (or "list my Martekio spaces") at any time. It returns every space you have access to and marks which one is the default.
The assistant passes the brand you name to each tool, and Martekio checks it against your own access on every call, so you can move between your brands freely but never see one you are not entitled to.