See why AI recommends your competitors — not you.
CitMark checks ChatGPT, Perplexity, Gemini and Claude every week. It tells you — in plain English — whether your site gets mentioned, gets cited, or gets skipped. Then it tells you exactly what to fix.
Founding pricing — 50% off for 12 months — is limited to the first 50 founding members. Create your free account — no card required. See our Privacy Policy.
Engines as run: ChatGPT gpt-6-luna · Perplexity perplexity/sonar · Gemini gemini-3.8-flash · Claude claude-haiku-5-5 · Grok grok-4.20-0309-non-reasoning · DeepSeek deepseek-chat · Mistral mistral-large-2512. Question set v2 — the same 3 buyer questions as the free check.
Free check
Check your llms.txt — right now, for real
Enter your domain — we'll fetch your llms.txt, the small file that tells AI what your site is about, and grade it on the spot. No signup, no email. Done in seconds. The full visibility check — whether those engines mention you at all — runs inside your account, and accounts are open now.
If this file is missing or thin, ChatGPT, Perplexity, Gemini and Claude may have trouble understanding what you sell — and who to recommend. Let's see how yours looks.
Sample report
…and this is what lands in your inbox every Monday
We ran CitMark for real on Oct 28, 2026 — eight brands, all seven engines, the same 3 buyer questions as the free check. Every score below is the actual result, low ones included.
| Brand | ChatGPT | Perplexity | Gemini | Claude | Grok | DeepSeek | Mistral | Score |
|---|---|---|---|---|---|---|---|---|
| Good Moleculesgoodmolecules.com | 1/3 | 1/3 | 2/3 | 1/3 | 3/3 | 3/3 | 3/3 | 66.795% CI 45.4–82.8 |
| Floydfloydhome.com | 2/3 | 1/3 | 2/3 | 1/3 | 0/3 | 1/3 | 1/3 | 38.195% CI 20.8–59.1 |
| Dusen Dusendusendusen.com | 2/3 | 0/3 | 1/3 | 0/3 | 0/3 | 2/3 | 1/3 | 28.695% CI 13.8–50.0 |
| Kotnkotn.com | 1/3 | 0/3 | 1/3 | 0/3 | 0/3 | 0/3 | 0/3 | 9.595% CI 2.7–28.9 |
| Wild Onewildone.com | 0/3 | 0/3 | 1/3 | 1/3 | 0/3 | 0/3 | 0/3 | 9.595% CI 2.7–28.9 |
| Cocokindcocokind.com | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0.095% CI 0.0–15.5 |
| Moon Juicemoonjuice.com | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0.095% CI 0.0–15.5 |
| CitMarkcitmark.com | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0.095% CI 0.0–15.5 |
Sample report — a real CitMark run on Oct 28, 2026.
Each engine cell shows in how many of the 3 buyer questions that engine named the brand. The score is the share of all 21 answers (3 questions × 7 engines) that named it, with the 95% confidence range. Engines as run: ChatGPT (gpt-6-luna), Perplexity (perplexity/sonar), Gemini (gemini-3.8-flash), Claude (claude-haiku-5-5), Grok (grok-4.20-0309-non-reasoning), DeepSeek (deepseek-chat), Mistral (mistral-large-2512).
Moon Juice scores 0 on the buyer questions — yet when we asked for alternatives to AG1, 4 of the 7 engines named it. AI knows the brand; buyer questions don't surface it. And CitMark is our own site: it is new.
That zero is real — we score CitMark with the same questions and the same ruler as every other brand in this table, and we're publishing the fixes as we make them.
Independent sample. CitMark is not affiliated with or endorsed by the brands shown.
Scores change week to week and vary by brand.
— Synced every price on this site to the current plans, after older prices were found still live on some pages. Engines repeat what pages say, so a stale price is a wrong fact an AI answer can pass on.
— Published llms.txt, a small file that tells AI systems what this site is and where its key pages are. Our own row in the table is 0 of 21; being findable and readable is the first fix.
— Repaired the /security page after a stylesheet error left parts of it rendering without their styles. A page that doesn't render as written is harder for engines — and people — to read and quote.
Want this report for your own site?
Create a free account — founding members lock in 50% off for 12 months. First 50 members only.
How we check
No black box — here's exactly what we do
Anyone can show you a score. This is how ours is produced, so you can judge for yourself whether it's worth trusting.
Every week we ask your tracked buyer questions — the things your customers type into AI right before they buy — in ChatGPT, Perplexity, Gemini and Claude.
For every answer we record whether your business is mentioned, whether your pages are cited as sources, and which competitors appear instead of you.
We grade your llms.txt on five things: it exists and isn't empty, it has a main heading, it's organised into sections, it links your pages with descriptions, and it covers the essentials — products, docs, API, contact.
How it works
Set it up once. Then it just runs.
Add your domain
Tell us your site and your competitors — every plan includes competitor tracking. We suggest the questions your buyers actually ask AI — you approve or edit the list. No code, no analytics access, nothing to install.
We ask the engines, weekly
Every week we put your buyer questions to ChatGPT, Perplexity, Gemini and Claude, and record who gets mentioned, who gets cited, and which pages the answers link to.
You get fixes, not charts
A Monday digest: score changes, mentions won and lost, competitor moves — and a short list of plain-English fixes ranked by impact, so you know exactly what to change next.
What gets monitored
Everything that decides whether AI recommends you
01Brand mentions
Does the answer name your business at all? We track mention rate per engine, per question, week over week — the earliest warning sign when you start slipping.
02Citations & source links
Mentions are nice; citations send traffic. We record exactly which of your pages each engine links to as a source, and which competitor pages beat yours.
03Competitor share of voice
See your slice of AI answers next to up to five competitors — and get alerted when a rival starts winning questions that used to be yours.
Plus: llms.txt health monitoring, and a plain-English fix-it brief in every Monday digest — the actions, not just the scores.
Why now
Search is moving into AI answers. Most small sites can't see whether they're in them.
Your customers increasingly get recommendations from ChatGPT, Perplexity, Gemini and Claude instead of a list of blue links.
If AI never mentions you, you don't exist in that conversation — and today, almost no small site can even see whether it's happening.
Enterprise brands buy Profound or AthenaHQ. Everyone else gets one-off free checkers that generate a file once and never look again.
CitMark is the missing middle: ongoing monitoring and plain-English fixes at a price a small site can justify.
Sources: llmstxt.org · Originality.ai, llms.txt adoption tracking · Gartner, Market Guide for Answer Engine Visibility Tools (Mar 2026)
The toolkit
Tools that turn your score into fixes.
Impact tracker
Did it work? See what changed after each fix — with noise removed.
Page checkup
Why AI skips your pages — and the brief to fix it.
Win citations
Sites that cite your rivals but not you — with outreach drafts.
AI readiness
Can AI read your site? A technical check, plus a one-line brief for your developer.
llms.txt generator
A small file that tells AI what your site is about.
Fix a bad mention
Step-by-step when AI gets your facts wrong.
Answer history
See how AI's answers about you changed, word for word.
What AI says about you Coming soon
Not just whether AI mentions you — what it says, checked against your facts.
Question radar Coming soon
New questions worth tracking — from search data, not guesses.
MCP access (read-only) Coming soon
Ask your CitMark data inside ChatGPT or Claude.
Pricing
Priced for the sites the enterprise tools ignore
Four plans, in USD, cancel anytime. Founding pricing — half price — is limited to the first 50 founding members, and lasts for 12 months, then renews at the standard price.
Starter
You're starting out and need to know where you stand in AI answers.
Founding price — locked for 12 months.
After your first 12 months, this plan renews at the standard $35/mo.
- 1 website
- 30 tracked buyer questions
- ChatGPT, Perplexity, Gemini & Claude
- Weekly checks + Monday email digest
- Mention & citation tracking
- Competitor tracking — up to 19 rivals
- Alerts when AI turns negative about your brand
- All nine tools included · depth grows with your plan
- 1 seat
Free account · no card required
Growth
You're actively competing for AI recommendations in your category.
Founding price — locked for 12 months.
After your first 12 months, this plan renews at the standard $70/mo.
- 1 website
- 100 tracked buyer questions
- Everything in Starter
- 5 AI engines — adds Grok
- Competitor tracking — up to 20 rivals
- Share-of-voice reporting & rival alerts
- Fix-it briefs ranked by impact
- Alerts when your score really moves — noise doesn't count
- Everything in Starter, with deeper tool coverage
- 3 seats
Free account · no card required
Pro
Your sales live on AI answers — you can't afford to miss a drop, a rival's move, or a negative mention.
Founding price — locked for 12 months.
After your first 12 months, this plan renews at the standard $169/mo.
- 1 website
- 300 tracked buyer questions
- Everything in Growth
- 6 AI engines — adds DeepSeek
- Competitor tracking — up to 20 rivals
- Fix-it briefs ranked by impact
- Alerts when a question drops you, or a rival overtakes you
- Everything in Growth, with deeper coverage across all nine tools
- 5 seats
Free account · no card required
Teams
You run several brands — or report on them to clients.
Coming soon
After the first 12 months, Teams renews at the standard $300/mo.
- 10 websites
- Everything in Pro
- 7 AI engines — adds Mistral
- White-label reports
- Every alert, across all your sites, in one roll-up
- Everything in Pro
- 20 seats — read-only seats don't count toward the limit
Launch prices: Starter $25/mo, Growth $49/mo, Pro $129/mo, Teams $199/mo.
Standard prices: Starter $35/mo, Growth $70/mo, Pro $169/mo, Teams $300/mo.
Founding pricing is limited to the first 50 founding members, applies to the Starter, Growth and Pro plans, and lasts for 12 months, then renews at the standard price.
No card required to create an account.
And if your first report finds nothing useful, we'll refund you — no questions asked.
FAQ
Straight answers
Still stuck?
Write to [email protected] — we answer by what your issue is, never by what you pay.
Each week we put your approved list of buyer questions — the things customers ask before buying — to ChatGPT, Perplexity, Gemini and Claude. We record whether your brand is mentioned, whether your pages are cited as sources, and how competitors did on the same questions. Weekly sampling keeps results meaningful and costs low; AI answers shift too slowly for daily checks to tell you more.
ChatGPT, Perplexity, Gemini and Claude at launch — the four engines your buyers are most likely to ask. We'll add more as buyer behaviour shifts.
No. No code snippet, no plugin, no Google Analytics or Search Console access. You give us your domain and your competitors; we observe the public answers the engines give to those questions. Works with any platform — Shopify, Wix, WordPress or custom: if you can type your domain, you're set.
They're built — and priced — for enterprise brand teams. CitMark is built for the long tail they don't serve: small sites, bloggers, indie makers and local businesses, from $25/mo, with fixes written in plain English instead of dashboards built for analysts. Free tools like Yoast's or Firecrawl's generators create an llms.txt once; they don't monitor anything week to week.
SEO tools optimise for Google rankings — but AI answers don't show rankings, they cite sources. You can rank #1 and still never be mentioned when a buyer asks ChatGPT. CitMark monitors the thing SEO tools can't see: whether AI actually names and cites you. Same for Google Search Console: it shows how Google sees your site; CitMark shows how AI engines see it.
Then you shouldn't pay for it. If your first report finds nothing useful, we'll refund you — no questions asked. Just email us and it's done.
Find out what AI says about you — before your competitors do.
Accounts are open now. Create yours free — founding pricing at 50% off for 12 months is limited to the first 50 founding members.
Or create your account on the sign-up page →
Questions first? Email [email protected].