Frequently asked questions

AI search tracking, plus a free tool directory. The paid product measures how often ChatGPT, Claude, Gemini and Perplexity name your brand when buyers ask what to use, explains why the answers pick your rivals instead, and gives you the page to publish to change that. Alongside it we keep a hand-written index of AI and SEO tools and a set of free utilities, both free to use with no account.

We ask each engine the buying questions your customers ask, on the same surface a customer would get: the current model with live web search switched on, not a cheaper model answering from memory. Every question runs twice per engine per sweep, because AI answers vary run to run and one sample cannot separate a real change from noise. Every answer is stored in full, so any number we show can be traced back to the text it came from.

Some difference is normal: assistants personalise, and the same question asked twice gives different lists, which is exactly why we sample twice and report consistency rather than a single result. If the difference is large and persistent, tell us the question and we will look, because it usually means we are measuring the wrong surface, and that is a bug worth fixing.

A visibility score per engine and the trend behind it, the rival quotes that justify the brands beating you, a ranked roadmap of what to ship, a written draft of the page for the question you are losing most, the third-party pages the answers cite and whether you appear on them, and an alert when you drop or a competitor gains ground.

Read it first. The draft follows our house style and is built from your own measured answers, but it is written by a model, so any figure it cannot source is left as a bracketed placeholder like [your price] for you to fill in. If a number slips through that we cannot trace to your answers, we flag it in red and tell you to check it. Treat the draft as a strong first pass from a writer who has read all your data, not as finished copy.

Every quote shown to you is checked against the stored answer it came from, and anything that cannot be traced back is dropped rather than shown. We do not sell prompt volume estimates, because no AI engine publishes how often a question is asked and every figure on the market is modelled. Where we cannot measure something, we say so instead of estimating it.

Rank trackers watch the ten blue links. We watch what an assistant actually answers when someone asks what to buy, which is a different result set with different winners. If your buyers open ChatGPT first, that is the ranking deciding whether you get considered at all.

Starter is $29 a month, Pro $49, Scale $149, and every plan starts with a 3-day trial. The trial takes a card and charges nothing until it ends, so cancelling before then costs you nothing. Plans differ by how many brands and questions you track, which engines run, and how much history you keep. Every plan runs weekly. There are also two optional add-ons on any plan: advanced answer models, which run the exact models chatgpt.com and claude.ai serve, from $59 a month; and deeper sampling, which asks every question four times per engine instead of twice, from $29 a month.

Yes. Every brand has a share link: a read-only page with the score, who the answers name instead, and the work queued up. It is a snapshot, so it will not change under whoever you sent it to, and you can turn it off in one click. Your stored answers and the rest of the dashboard stay private.

We add tools by hand, write the descriptions ourselves, and rank by how useful each tool is for its job, weighted by editorial judgment. Nobody can pay for a listing, a ranking, or a verdict. Being listed is free: use the contact form and choose Submit a tool.

Subscriptions to the tracking product, and some outbound directory links are affiliate links, which means we may earn a commission if you buy after clicking. Affiliate status never changes what we list, how we rank, or what a verdict says; plenty of our top picks pay us nothing.

We check pricing when we add a tool and refresh it regularly, but vendors change plans without telling us. Treat our pricing as a reliable snapshot and confirm on the vendor's site before buying. Each pricing page shows when it was last checked.

Yes, with fair-use rate limits so nobody drains them. Some tools unlock their full output (like the visibility checker's complete report or the keyword CSV export) with your email, which also subscribes you to the weekly AI Search Brief. Unsubscribe anytime; the tools keep working.

Visibility checker inputs and reports are stored for up to 90 days so results load instantly for repeat checks. Text you paste into the prompt optimizer and AI content checker is processed to generate your result and not saved. The details live in our privacy policy.

A weekly email about AI search: new tools worth knowing, changes in how AI engines cite sources, and practical ways to keep your brand visible. One email a week, no filler, unsubscribe anytime.