Pricing
Prices in US dollars. Every tier runs the same frozen buyer questions, so results are comparable over time. No accounts, no calls. Pay once, or subscribe when you want re-tests.
What we sample, exactly. Samples from the OpenAI web-search API (model shown in the report) and the Perplexity Sonar API — not the consumer ChatGPT or Perplexity apps, whose answers also depend on memory, location and personalisation. Every report prints the model id, the search tool and the run time, so you can tell one from the other. How we measure.
See the shape of it, free. Up to 10 questions × 3, one engine, no account and no card: who gets recommended (share of sampled answers), plus whether your homepage is reachable. Not included: the source ledger (that one needs an email), per-page reachability, and the fix list — those start at the Check. With 3 runs the numbers come as a range over 3 runs, not a confidence interval.
Run the free snapshotThe one report to start with. Up to 40 buyer questions × 3 on OpenAI + Perplexity = up to 240 answers in total, every prompt verbatim. You get who gets recommended (share of sampled answers), a page-verified ledger of what each engine cited instead, reachability for 6 of your pages, and a fix list. Usually back in about 20 minutes. One-time — nothing renews.
Buy the Check — $29 Read a real Check reportWhen you already know what's missing. Everything in the Check on OpenAI + Perplexity, plus the artefacts: technical diffs, profile copy for the review sites these engines read, correction emails, and a day-30 re-test. Up to 5 first drafts written only from the facts you enter — and none at all where your own pages already answer the question, in which case you get that page named and what to strengthen instead. All unpublished until you approve. $29 credited.
Buy the Fix Pack — $99 See a real Fix PackFor sites that change every month. Everything in the Fix Pack, then a re-test every 14 days on the same frozen questions, a rolling fix queue, a "what changed after this fix" tag on every approved item, and reachability alerts. $99 credited against your first payment.
Subscribe — $149/mo or $1,188 USD/yearThe ladder credits forward: the Check's $29 comes off the Fix Pack, and the Fix Pack's $99 comes off your first subscription payment. Buying the cheap thing first never costs you extra.
$399 USD/mo for 5 client sites. Every site at subscription depth on OpenAI + Perplexity: white-label PDFs with your client's name on them, share links you can send instead of building a slide, capped pitch snapshots for prospects you have not won yet, and a JSON API with webhooks for your own dashboard. You start it yourself — no demo call, no contract, no minimum term.
Start the Agency planA re-test can only see a change large enough for this design to resolve. At 40 questions × 3 repeats, the Fix Pack's day-30 comparison has roughly 80% power for a change of about 20 points. At 80 questions × 3 repeats, the subscription's 14-day comparison has roughly 80% power for a change of about 15 points, and each Agency site re-tests at 80 × 3 on the same design — about 15 points, on the same 14-day cadence. It is not a cutoff: smaller changes can sometimes be detected and larger ones are not guaranteed to be, but this design is mainly able to see large, directionally consistent moves however often it runs.
Which one is for me
- You have never measured this. Run the free snapshot, then buy the Check. The snapshot tells you who is being named; the Check tells you why, and what to do.
- You already know you are invisible and want the work done. Go straight to the Fix Pack — it contains the Check, and the artefacts are the point.
- Your site or your category changes every month. The subscription, because a one-off reading of a moving target ages fast and the re-tests use the same frozen questions.
- You report to clients. The Agency plan: five sites, white-label PDFs, share links and an API, with no call to book.
- You want to check our arithmetic first. Read the method, then read a real report. Both are public and neither asks for an email.
Tier by tier
Scroll for more →
| Tier | Price | Sample | What you end up with | Re-tests |
|---|---|---|---|---|
| Free snapshot | $0 | Up to 10 questions × 3, one engine | Who gets recommended, and whether your homepage is reachable | None |
| AI Visibility Check | $29 USD, once | Up to 40 × 3 on OpenAI + Perplexity | The full source ledger, per-page reachability for 6 pages, a fix list and a shareable report | None |
| Fix Pack | $99 USD, once ($29 credited) | Up to 40 × 3 on OpenAI + Perplexity | The Check, plus diffs, drafts, profile copy and correction emails — all unpublished until you approve | One, on day 30 |
| Subscription | $149 USD/mo, or $1,188 USD/year | Up to 80 × 3 on OpenAI + Perplexity | The Fix Pack, a rolling fix queue, per-fix attribution and reachability alerts | Every 14 days |
| Agency | $399 USD/mo, 5 sites | Up to 80 × 3 per site on OpenAI + Perplexity | Subscription depth per site, white-label PDFs, client share links, capped pitch snapshots, a JSON API with webhooks | Every 14 days |
Questions people ask before buying
What do I actually get for $29?
One report on your own domain, usually back in about 20 minutes. Up to 40 buyer questions in your category, asked 3 times each through the search APIs named in the box at the top of this page, with every prompt printed verbatim. In it: who gets recommended (share of sampled answers) for you and for every brand the engines actually named; a page-verified ledger of which pages each engine cited; reachability cards for 6 of your pages, tested from a crawler's side and a browser's side; and a fix list ordered by what those pages are missing. It is a one-time purchase — nothing renews, and no account is created. Read a real one before you pay.
How do I know the numbers are real?
Because you can check them. Every prompt is printed exactly as it was asked, and the report names the model id, the search tool and the run time it used, so you can re-ask any question yourself. Rates come with a question-resampling sensitivity range rather than a single number, so you can see signal from noise. And we say what we sampled, in the box at the top of this page — the search APIs, never the consumer assistants, whose answers also depend on memory, location and personalisation. The whole procedure is published at methodology, including the parts we refuse to score and why.
What do I do with it on Monday morning?
Three things, in this order. The reachability cards name the exact pages a crawler could not fetch, and why — a robots rule, a firewall, content that only exists after JavaScript. That is a ticket for whoever owns your site. The source ledger names the third-party pages the engines cited instead of you, usually a listicle or a directory, and those are pages you can go and correct. The fix list orders your own questions by what is missing from the closest page: a price, a number, a comparison, steps. If you would rather have that turned into finished artefacts — diffs, drafts, profile copy, correction emails — the $99 Fix Pack does it, and your $29 comes off.
Why $29, when the snapshot is free?
The free snapshot is real and it is not a teaser: 10 questions, 3 runs, one engine, no account, no card. It tells you who is getting recommended. It does not tell you which pages were cited, whether the engines can read yours, or what to fix — those need many times the API calls, and the $29 pays for them. Run the free one first if you want to see the shape of it; the questions are generated the same way.
What happens if the scan doesn't deliver?
Report usually within 20 minutes, at most 24 hours once a scan can actually run. If something stops it running at all — a login wall, a wrong domain, an engine outage on our side — we tell you what to fix and re-run it free rather than starting the clock. If it can run and we still cannot produce a report with the agreed questions and engines within 24 hours of that, you are refunded in full. If a tier promises two engines and only one was available, we say so before you pay, and the missing engine is re-run free first, or refunded. What is not a refund reason: the numbers not going up. We never promised they would. Full refund policy.
Do I need an account, a call, or a contract?
No, no, and no. Nothing here creates an account: you pay, a one-time start link arrives by email, and that link runs your scan. There is no demo call to book — Kovmere is built and run by one person, Claire Wang, and the whole product is software you can inspect. The $29 Check and the $99 Fix Pack are one-time purchases; only the subscription and agency plans recur, and both cancel from a link in your own receipt.
What this is not
- Not a promise about rankings or citations. We report what the engines answered on the day we asked, with the range that goes with it. Engines change. Anyone selling you a guaranteed mention is not measuring one.
- Not schema, llms.txt, IndexNow or FAQ re-formatting. Controlled studies show no measurable effect on citations, so we don't grade them and we won't sell you work on them.
- Not bulk AI pages. A Fix Pack writes up to 5 pages, from facts you enter, on your domain, published only after you approve — and none at all where your own page already answers the question.
- Not the consumer apps. We sample the search APIs named at the top of this page, never the consumer assistants, whose answers also depend on memory, location and personalisation.
- Not a human reading your site. Nothing here is edited by a person or published on your behalf. One person builds and runs Kovmere; the rest is software, and the method is published.
- Not a subscription unless you choose one. The $29 Check and the $99 Fix Pack are one-time. Nothing renews.
Refunds: if a paid scan fails to complete, you get a new token or your money back. See refund policy. Charges appear as KOVMERE on your card statement.