Enter your brand and keywords for a 30-second visibility checkup. No login.
Keywords are matched against public industry queries only and won't be used in private searches.
Privacy: your IP is HMAC-pseudonymised for abuse prevention and never published. Results are approximate.
You give one brand name
Industry keywords are optional — they sharpen the prompts, and we infer a category if you skip them.
We ask 5 engines category questions
We ask "which X are worth considering", never "how good is X" — a question that names you can't measure visibility.
You get a tier per engine, plus a score
The result gets its own link you can send to a colleague — no account needed on their side.
Half Western, half Chinese. A tool that covers only one side quietly implies your other side looks the same — the most expensive kind of wrong.
Every engine-by-prompt cell lands in one of the three tiers below, then they're weighted per engine into a 0–100 score.
| Check | Weight | Why this much |
|---|---|---|
| Named in the answer body | 0.5 | The AI says your brand name in the answer itself — the highest tier a free checkup can award. |
| Only in the cited sources | 0.2 | Not in the prose, but your site appears in the citation list. Present, though most users won't see it. |
| Absent | 0 | This engine doesn't know you for this question. |
There is a fourth tier — "recommended first" — that a free checkup cannot award: one day's snapshot isn't enough to judge it reliably, so it exists only in continuous monitoring. That's why checkup scores skew conservative by design, not by accident. Also: cells that haven't run yet are marked "generating", stay out of the denominator and cannot drag your score down. If none have run, we label it "not enough data" rather than handing you a 0 that looks real.
It can tell you
It can't tell you
No account, no payment. Enter a brand name and it runs; the result gets its own shareable link. A free account adds history and 3 checkups a month.
Because a checkup only scores up to "mentioned" — it can't award the top "recommended first" tier, which needs continuous monitoring to judge reliably and only exists in the paid plan. So checkup scores skew conservative by design: they answer "does AI know you", not "how strongly does AI recommend you". A lower checkup score than a monitoring score is expected.
That engine-and-prompt combination has no usable public answer yet for the day. Cells that haven't run are excluded from the denominator — they can't drag your score down. If none have run, we label it "not enough data" rather than handing you a 0 that looks like a real score.
Category questions — "which <your category> are worth considering" — not "how good is <your brand>". If the question names you, the AI will almost always mention you, and you end up measuring your own phrasing instead of your visibility. We learned this the hard way: an early prompt set named the brands, and for one of them 288 of 296 cells had to be discarded.
The checkup runs five: ChatGPT, Kimi, DeepSeek (Volcengine), Gemini and Doubao — half Western, half Chinese, which is the main thing that separates us from tools that only do one side. Paid monitoring covers more.
No. We only read public AI answers. No tags, no visitor tracking, no code changes on your side — put another way, we can't see your traffic, and we don't want to.
All standalone, all free to finish without an account.
Diagnose
Static GEO-readiness check for any page, no login
Benchmarks
Public, anonymized industry percentiles
Embed widget
One line of code to embed the checkup
