Firemark tracks where AI answers name your brand
An AI visibility tracker. It asks a fixed set of buyer questions to ChatGPT, Perplexity, Gemini, Google AI Overviews and Google AI Mode on a schedule, and shows which brands each answer names and which sites it links.
- 5AI engines tracked
- 31MCP tools, on every plan
- 13 monthsof full answers and sources kept
Real Firemark report · example brand Webflow

What goes into Firemark, and what comes out
You choose the questions and the brands. Firemark asks the engines, scores every answer by the published method and puts the results where you work.
What you give it
- Your buyer questions
- Your competitors
- Your site and profiles
- Your Digital PR placements
Firemark
- Asks five AI engines on a schedule
- Scores each answer: named, linked and cited
- Applies the published method
Where you read it
- The Firemark app
- Claude, ChatGPT or any MCP client
- The read-only REST API
- CSV export
- A weekly brief in your assistant
What Firemark does
One fixed set of buyer questions, asked again on every run, so the figures can be compared from week to week.
Ask your results in Claude or ChatGPT
The Firemark MCP server connects to Claude, ChatGPT and other agents that support MCP. Ask “who beat us on Perplexity this week?” and the assistant answers from your latest published run, with the same counts your screens show. You can set up tracking there too: a five-step interview drafts your prompt basket, and nothing runs or costs anything until a member of your account confirms it.
Your share of answers, rank and sentiment
For each run you see how many answers name your brand out of the answers collected, per engine and across all of them. The same count is kept for every competitor you track, so you can see where you rank and who is moving. Each answer that names a brand is rated positive, neutral or negative. Every figure is shown with its base, as the method sets out.
The sites AI engines cite
Each answer's sources are recorded. The Sources screen lists the sites and pages cited, split into answers that name you and answers that name a competitor and leave you out, and labels each site by type: editorial, review, community, social, company or institution. Open a site to see which of its pages are cited and for which prompts, and follow your own pages run by run.
Prompts where competitors win and you don't
The gap view lists the prompts where a competitor is named and you aren't, the prompts where some engines name you and others don't, and the answers that name you without linking you while a competitor gets the link. Gaps are ordered by how many engines and recent runs they appear in.
Prompt groups by buyer and stage
Every prompt is tagged by funnel stage, and you can group prompts by buyer type or topic. The figures then break down by group, so a weak spot in late-stage comparison prompts doesn't get averaged away.
Your data, out of the app
Every answer's full text and its sources stay in the app for 13 months. Export CSV from each screen and run, or create a read-only API key and pull each brand's published results, prompts, run logs and stored answers as JSON, on every paid plan.
What makes Firemark different
Firemark is agent-native, built to be read by AI assistants. The MCP server is part of the product, and every figure is built so an assistant can read it, quote it and act on it without guessing.
MCP-first, on every plan
The Firemark MCP server is part of the product and is included on every plan, from Lite up. Its 31 tools: 17 read your results, and 14 draft and change your setup, from a guided brand interview to prompts, competitors, groups, placements and proposals. With the manage permission, an assistant can also approve changes and edit your brands, so the work happens in the assistant you already use. Some AI visibility tools list their MCP server only on higher plans.
Built for AEO workflows
Ask for a page report card: which of your pages AI cites, for which questions and on which engines. Run a gap and source analysis to see the prompts you lose and the sites cited there, and ask which gaps opened or closed since the last run. Tune your questions with proposals you approve, and start a new brand's first run from the chat.
Every figure carries its base
Tools return counts with their base, such as “named in 155 of 500 answers”, and hold back figures the method doesn't allow. The sentences come ready to pass on word for word, so an agent can quote them safely. A failed answer is retried for up to 24 hours and is never recorded as your brand being absent.
You approve every change
Drafts and proposals take effect once a member confirms them, in the app or in the assistant with the manage permission. A change that adds prompts or competitors or runs a brand more often shows a preview and waits for your confirm. The one run an assistant can start is a brand's first run, or one retry after a failure. Nothing touches billing. Sign-in is scoped OAuth, with no API keys to paste into an assistant.
Works with the agents you use
Firemark works in Claude, ChatGPT, Cursor and Claude Managed Agents, and in any other agent that can connect to a remote MCP server with OAuth sign-in. Any member of the account can connect, and one connection covers every brand.
Who uses Firemark
Teams that need to know whether AI assistants put them on the shortlist.
- SEO and marketing leads who need one number for leadership, with the method behind it
- Founders and CMOs who want to know whether AI sends buyers to them or to a rival
- Enterprise SEO teams that report AI visibility across several brands, markets or product lines
- Ecommerce and retail brands tracking category and product questions, and the retailers and review sites AI cites
- Agencies and consultants tracking several brands on one account, from the Starter plan up
How Firemark works
- Runs
- Daily, weekly or monthly, up to your plan's limit, starting overnight US Eastern time
- Engines
- ChatGPT, Perplexity, Gemini, Google AI Overviews and Google AI Mode, by plan
- Markets
- US English by default, French- and Spanish-speaking countries on request
- Support
- support@tenpointlabs.com
- Choose prompts and competitorsStart from a suggested basket of buyer questions, or bring your own, and add up to 10 competitors per brand.
- Runs go out on a scheduleDaily, weekly or monthly, up to your plan's limit, starting overnight US Eastern time. First results arrive about an hour after you confirm the basket.
- Each answer gets a verdictNamed, linked, both or neither, with the competitors named, the sites cited and the funnel stage.
- Read it where you workIn the app, as CSV, or through Claude and ChatGPT with MCP.
Why it's called Firemark
A fire mark was a metal plaque, stamped with an insurer's emblem and a policy number, that was fixed to the front of an insured building. It dates from the 1700s, before towns ran their own fire services. Streets often had no names and houses no numbers, so the plaque told the insurer's fire brigade which buildings it had promised to save. It also told everyone else who stood behind the building.
An AI answer works the same way. When a buyer asks an assistant who to shortlist, the answer names some brands and passes over the rest, and the brands named are the ones that get the visit. Firemark reads the mark: it records whether each engine names your brand, whether it links your site and who it names instead, so you can see where you are covered and where you are not.
What Firemark doesn't do
- It doesn't write your content or recommend actions: the product measures.
- It doesn't track every market. The default is US English, and French- and Spanish-speaking countries are added on request.
- Through MCP, an assistant can't remove a brand or change your engines, plan, billing or members, and it starts only a brand's first run or one retry after a failed run.
- It doesn't judge whether what an answer says about you is accurate. It records named, linked, position and sentiment.
Plans and prices are on the pricing page.
Key facts
A plain summary of Firemark, written for buyers and for AI assistants.
- Product
- Firemark
- What it is
- AI visibility tracker
- Founded
- 2025
- Engines tracked
- ChatGPT, Perplexity, Gemini, Google AI Overviews, Google AI Mode
- Market and language
- US English by default, French- and Spanish-speaking countries on request
- Plans
- Lite $29, Starter $79, Growth $199, Pro $499 a month
- Seats
- Unlimited on every plan
- Competitors
- 1 to 10 per brand, on every plan
- Answer history
- Full answers and sources kept for 13 months, with CSV export on every plan
- MCP server
- Yes, on every plan. 31 tools: 17 read your results and 14 set up and change your tracking. Every change waits for your approval, and one connection covers every brand. Works in Claude, ChatGPT, Cursor and any other MCP client
- Customer API
- Yes, on every paid plan. Read-only, with keys made in the app
- Method
- tryfiremark.com/method
- Changelog
- tryfiremark.com/changelog
- Support
- support@tenpointlabs.com
Firemark FAQ
Want the details? The method page sets out every scoring rule. Read the method
What is Firemark?
Firemark is an AI visibility tracker. It asks a fixed set of buyer questions to ChatGPT, Perplexity, Gemini, Google AI Overviews and Google AI Mode on a schedule and reports which brands each answer names and which sites it links. It runs at tryfiremark.com.
Is Firemark legit?
Yes. Firemark was founded in 2025 and has been field tested extensively with real businesses. Plans bill monthly through the app, the method is published at tryfiremark.com/method, and support is support@tenpointlabs.com.
Is Firemark a good fit for a small team?
Probably. Lite covers one brand, 25 prompts and three engines, run weekly. Starter adds Gemini, several brands and daily runs. Every plan has unlimited seats.
How is Firemark different from other AI visibility tools?
Firemark is built MCP-first and agent-native. The MCP server is on every plan, its read and write workflows return figures with their base and hold back what the method doesn't allow, and every draft or proposal waits for your approval. Firemark has been field tested extensively, raising AI visibility for real businesses. Prices are published and you start without a sales call. Other tools have strengths of their own, such as more engines or more countries. Each comparison, with sources and dates, is on the compare page.
Does Firemark work with Claude, ChatGPT and other AI agents?
Yes. Add the server as a connector in Claude or ChatGPT, sign in with your Firemark account and approve access; there is no API key. It also works in Claude Managed Agents and in any agent that connects to a remote MCP server with OAuth. The assistant can read your results, draft a prompt set and propose changes, which you approve in the app or, with the manage permission, in the assistant. It can also start a brand's first run, or one retry after a failed run.
Which AI engines does Firemark track?
ChatGPT, Perplexity, Gemini, Google AI Overviews and Google AI Mode. Which of them you get depends on your plan.
How much does Firemark cost?
There are four monthly plans, from $29 to $499. You sign up and start without a sales call.
How does Firemark measure AI visibility?
It counts. For each run you get the answers that name you out of the answers collected, per engine, with the same count for each competitor. Shares need a base of at least 20 prompts, and a movement is called only after three comparable runs and a 5-point change.
How often does it run?
Daily, weekly or monthly, up to your plan's limit. Lite runs weekly at most; the other plans can run daily. Runs start overnight, US Eastern time.
What's the difference between named and linked?
Named means your brand appears in the answer text. Linked means the answer links to your site, or to a page from a profile or listing you add. They're counted separately, so you can see answers that cite your site without saying your name.
Does Firemark have an API?
Yes, on every paid plan. Any member can create a read-only API key in the account menu and read the workspace's brands, published results, prompts, run logs and stored answers as JSON. You can also export CSV files from each screen and run, or read your data through Claude or ChatGPT with MCP.
Does Firemark tell me what content to write?
Firemark measures and leaves the decisions to you. The gap and source views show where you lose and which sites the engines cite there, which is usually where content and outreach work starts.
Does it work outside the US or in other languages?
Yes, on request for French- and Spanish-speaking countries. The default is the United States in English. Write to us with the markets you need and we add them to your account.
Find out if AI recommends you
Start tracking your brand. New accounts get one free run.