How the audit works.

The free snapshot asks 4 engines (ChatGPT, Claude, Gemini, Perplexity) 5 buyer questions once each, runs 14 checks on your site, and shows who each engine linked to: you, a rival, or nobody. The full audit asks all 5 engines (Grok included) 20 questions, repeats ChatGPT, Claude and Perplexity three times so the numbers settle, keeps every answer with its links, gives you a CSV of the whole matrix, and writes a fixes pack for your site: answer-first FAQ copy first, then the crawler, schema and llms.txt files.

The snapshot is free, one per domain per 30 days. The full audit is $19 excl. tax, paid once via Stripe (tax is added at checkout where applicable): USD 19 in most of the world, EUR 19 incl. VAT in the EU, GBP 16 incl. VAT in the UK. It covers 20 questions across five engines; ChatGPT, Claude and Perplexity run three times, Gemini and Grok once. There is no subscription and nothing recurring: monitoring tools cost $240 to $1,188 a year at the entry tier, and other one-off audits start at $249. What each charges, and when a one-off is enough, is on the cost page.

A snapshot takes 2-3 minutes; a full audit 10-15. The report page shows each step as it happens, and we email you the link when it is done.

Linked means the engine's answer included a link to your domain or a subdomain. Named means your brand appeared in the answer text without a link. Both are shown as a share of all the answers we collected. Linked is the one that sends visitors; named is the one that tells you the engine knows you exist.

AI engines give a different answer every time you ask. So we ask more than once and report how often each site was linked, not a single verdict. We also reach the engines through their official APIs with web search switched on, which is close to what a person sees in the app but not identical. Treat every figure as a frequency, and treat the full audit's repeated runs as the reliable version. The methodology page gives the error bars per tier.

From your site we identify your category and your competitors, then write the questions a buyer in that category would ask an AI assistant, in five shapes: best X for Y, A vs B, alternatives to C, is D worth it / how much does X cost, and how do I do the job. None of them name your product, because a buyer who doesn't know you yet wouldn't.

ChatGPT (OpenAI), Claude (Anthropic), Gemini (Google), Grok (xAI) and Perplexity, each through its official API with web search on. The free snapshot skips Grok; the full audit includes it. Exact model IDs and search backends are on the methodology page.

Files you paste, not advice, in order of expected effect: answer-first FAQ entries written for the questions you're losing or that nobody wins yet (the lever with evidence behind it), a robots.txt addition if you block AI crawlers, a JSON-LD block for Organization and Product/SoftwareApplication, and a generated llms.txt. A README says where each file goes and what to expect from it.

Yes. A snapshot only reads public pages and asks the engines public questions, so you can check a competitor or a prospect. The report link goes to the email you give us.

The free snapshot runs once per domain every 30 days; a repeat request inside that window opens the existing report. A full audit can be bought any time from a finished snapshot. Engines pick up site changes over days to weeks, so re-check after you have published the fixes, not the same afternoon.

If a full audit fails to complete, the payment is refunded automatically and you get an email saying so. If a completed audit is clearly wrong (wrong site read, off-category questions, missing fixes pack), tell us within 14 days and we re-run it or refund it. Details in the terms.

Ask it the questions a buyer would ask - "best X for Y", "alternatives to Z" - and look at which sites the answer links, not which names it mentions; then ask again, because the answer changes between runs. A free snapshot does this for you: 5 buyer questions to ChatGPT, Claude, Gemini and Perplexity, and a table of who each one linked. One chat is one sample; the snapshot is 20.

Monitoring tools that track AI citations continuously charge from USD 20 a month (Rankscale) and USD 29 (Otterly) up to EUR 85 (Peec AI) and USD 99 (Profound, Semrush) at the entry tier — USD 240 to USD 1,188 over a year, and more on the plans above it. Other one-off audits run from USD 249 (BizWhiz) to USD 5,000 (Avante). A one-off audit that runs the questions once and gives you the fixes costs USD 19 here, with a free 5-question snapshot before it. Pay monthly when you will watch the trend every week; pay once when you need the diagnosis and the fixes.

In order of effect: get onto the pages the engines already link for your questions (listicles, comparison posts, review sites); put a direct answer to each buyer question in the first sentence of a page, under that question as a heading; put your own numbers on the page; let the search crawlers in (OAI-SearchBot, Claude-SearchBot, PerplexityBot); keep the page dated and current. The evidence behind that order is in the guide; the full audit writes the answer-first copy for the questions you are losing.

We store the URL, the email you give us, the crawled facts about your public site, the questions, and the engine answers, so the report page keeps working. The report link is unlisted; anyone with the link can view it. Your IP address is stored only as a hash, for the daily limit. To have an audit deleted, use the contact form (topic: privacy request). Full privacy notice and terms.