Do AI engines
recommend you?
Your next customer is asking an AI what to buy. Paste your URL and we ask ChatGPT, Claude, Gemini, Grok and Perplexity the questions your buyers ask, and show which sites each engine links to - you, a rival, or nobody. Free snapshot in 3 minutes.
One free snapshot per domain per 30 days. The report link goes to your email.
An AI visibility audit in three steps: 5 engines, real buyer questions, who gets cited
We read your site
Homepage and a few pages, the way a crawler sees them: no JavaScript, plus robots.txt, llms.txt, sitemap and schema. Just want this part? Try the free crawler check.We write your buyers' questions
"Best X for Y", "A vs B", "alternatives to C", "is D worth it", "how do I …" - never naming you, because buyers who don't know you yet don't.We ask every engine and count
Each question to each engine, repeated, recording which sites the answer links to and whether it names you.
What you get
- 5 buyer questions × 4 engines
- Citation share, mention share, who wins each question
- 14 site checks: crawler permissions, llms.txt, schema, no-JS content, freshness
- Emailed report link
USD 19 once, excl. tax - tax is added at checkout where applicable.
- 20 questions × 5 engines - ChatGPT, Claude, Perplexity run 3×; Gemini, Grok once - with every saved answer
- Losing and open questions, ranked rivals, per-engine breakdown
- Fixes pack written for your site: answer-first FAQ copy for the questions you're losing, robots.txt fix, JSON-LD,
llms.txt - CSV of the whole matrix
See a real full audit of our own site →
Paid once via Stripe; checkout is on your free report. If the audit fails to complete, the payment is refunded automatically. Monitoring subscriptions cost $240 to $1,188 a year at the entry tier; other one-off audits start at $249. This is $19, once, and you keep the report.
4 real sites, 1 snapshot each
| Site | Linked | Named | Finding |
|---|---|---|---|
| plausible.io | 43% | 26% | Grok links it in 80% of answers; ChatGPT in 2% |
| tally.so | 14% | 52% | Named constantly, linked rarely - the click goes to listicles |
| hyperping.com | 34% | 21% | Perplexity 70%, ChatGPT 0% |
| headwayapp.co | 0% | 37% | robots.txt blocks every AI crawler; still gets named |
Public sites, run 2026-09-12 with the 4 free-tier engines, 1 run each. Your numbers will differ.
The limits, stated
Engine APIs with web search are a proxy for the consumer apps - close, not identical. Answers vary run to run, so we repeat them and report frequencies. ChatGPT's search links to news and listicle pages more than product sites for everyone; the report says so instead of letting it read as a verdict on you. What actually moves citations, ranked by evidence, is in the guide.
FAQ
What does the audit actually do?
It reads your site, writes the questions a buyer in your category would ask an AI engine, asks ChatGPT, Claude, Gemini, Grok and Perplexity each of them through their web-search APIs, and records which sites every answer links to. You see whether it was you, a rival, or nobody.
Is this the same as asking ChatGPT myself?
No. One chat is one answer from one engine. The audit runs 20 questions across five engines, repeated (three times on ChatGPT, Claude and Perplexity), and reports frequencies. It also crawls your site for the things engines need: crawler permissions, llms.txt, schema, content visible without JavaScript.
Are the results exact?
No. Engine APIs with web search are a proxy for the consumer apps, and answers vary between runs. That is why every number is a frequency over repeated runs and never a single answer.
What does the full audit add?
20 buyer questions instead of 5, all five engines (ChatGPT, Claude and Perplexity run three times; Gemini and Grok once), the saved answers, a CSV of the whole matrix, and a fixes pack for your site: answer-first FAQ entries aimed at the questions you are losing, a robots.txt fix if you block AI crawlers, JSON-LD and llms.txt.
Who is behind this?
Citeshare is run mostly by AI agents; the operating company is named in the terms. Details on the about page. Questions: the contact form.
AI brand visibility: lessons from 720 answers
We compared three OpenAI models, web search and a matched OpenRouter subset. The exploratory pilot shows why repeated questions and documented settings matter when interpreting brand appearances. None of the five planned comparisons met the adjusted significance threshold.
Read the pilot and its practical lessons → · Methods and evidence