What they advertise, against what the entry price actually buys

Every vendor in this category gates something, and gating is what tiers are for. The question worth asking is whether you can see the limit at the moment you pay.

Every row quotes the claim rather than paraphrasing it, names where the quote came from, and dates when it was read. Where a fact came from a review roundup instead of the vendor, the row says so and its confidence drops. Where we could not confirm something, it reads as unverified rather than being guessed in our favour.

We are in this ledger, and our row is not flattering. Our marketing names four engines and our cheapest paid plan measures one. That is the same shape as a row below, written in the same words. A comparison page that exempts its author is a competitor grading competitors, which is what this category is already full of.

SEOAIOThat is us

Stated where you buy

What the marketing says: 4 engines named across the marketing site: Perplexity, Google AI Mode and AI Overviews, ChatGPT and Claude.; seoaio.ai homepage and /how-we-measure, read 2026-07-31
What the entry price buys: The $99 Starter plan measures citations on 2 engines: Perplexity and Google AI Mode. The rest unlock at Pro. Our pricing page prints that on the Starter card, in those words, before you pay.; seoaio.ai/pricing, Starter card, and CITATION_ENGINES_BY_PLAN in our own source, read 2026-09-06 (re-audited; the previous hand-typed sentence had gone stale)

Source: This ledger includes its author on purpose. The gap is real; the disclosure is the difference. Audited against our own code 2026-07-31: the Starter engine limit is on the Starter card, and two Agency lines that were NOT disclosing properly were corrected the same day (a sampling-depth superlative that was true of one engine out of four, and a cohort-benchmarking line that did not name its five-site floor).

Scrunch AI

True, but not on the purchase path

What the marketing says: Five AI engines named in the marketing.; scrunchai.com marketing pages, read 2026-07-27
What the entry price buys: The Core plan, reported at $250/mo, tracks four: ChatGPT, Perplexity, Google AI Overviews and Copilot. Claude, Gemini, Meta AI, Grok and AI Mode require Enterprise. Hallucination detection is Enterprise too.; Third-party review roundups, 2026-07-27. Scrunch does not publish pricing on its own site, which is why this row cannot be first-party.

Perplexity Computer (AI SEO page)

True, but not on the purchase path

What the marketing says: “Track your AI visibility in minutes”: checks your brand across ChatGPT, Gemini and Copilot and builds a playbook automatically.; perplexity.ai/gen/computer/ai-seo, read 2026-09-06 (a noindex landing page for their general agent)
What the entry price buys: Not a product and not priced on the page: it is Perplexity Computer, bundled into Pro ($20/mo) and Max ($200/mo), metered in credits the page never mentions. The roster it checks does not include Perplexity itself, nor Google AI Mode, nor Claude. The visibility score has no stated sample count or method.; Third-party pricing explainers, 2026-09-06; the page itself carries no pricing, no methodology and no sample output. Roster read from the page.

Opinly

Not what the entry price buys

What the marketing says: “Where your brand gets mentioned by AI”, in their advertising.; Opinly ad creative, read 2026-07-28
What the entry price buys: Their /platform page enumerates six toolkits and names ZERO AI engines anywhere on it. The only engine list on the entire site sits inside a blog post: ChatGPT, Google AI Overviews, Gemini, Perplexity and Claude. There is a real LLM module underneath, seven MCP tools deep, and the product pages do not sell it.; opinly.ai/platform and /blog/llm-content-optimization, read 2026-07-28

Source: teardown-opinly-2026-07-28

Ahrefs Brand Radar

True, but not on the purchase path

What the marketing says: AI-platform tracking presented as part of the Ahrefs product line, on a product page with no price on it.; ahrefs.com/brand-radar, read 2026-07-30
What the entry price buys: $199/mo per platform, or $699/mo for all platforms, charged ON TOP of the base plan. A customer on a $249/mo base plan sees that only in the in-app upgrade modal, after subscribing.; Ahrefs in-app upgrade modal, seen 2026-07-30 by a customer on a $249/mo base plan

Source: teardown-ahrefs-brand-radar-2026-07-30

Ahrefs custom prompts

Stated where you buy

What the marketing says: 2,500 custom prompt checks included at the top tier.; ahrefs.com/custom-prompts, read 2026-07-30
What the entry price buys: The unit is compound: one check per prompt, per platform, per location, per day. Ten prompts across six platforms in three locations, checked daily, consumes 180 checks a day, so the included allowance is about a fortnight of that configuration.; ahrefs.com/custom-prompts metering description, read 2026-07-30

Source: teardown-ahrefs-brand-radar-2026-07-30. The metering rule is stated plainly on their page; the arithmetic is ours.

PressWhizz (AI Visibility Seeding)

Not what the entry price buys

What the marketing says: “Become THE go-to answer for your customers’ queries in all LLMs (ChatGPT, Claude, etc.) in couple weeks”, beside “only 4 slots left for this quarter”.; presswhizz.com/services/ai-seo-company/, the indexed page for the offer (index, follow; listed in the sitemap), raw HTML read 2026-09-28
What the entry price buys: Tier 1 is $7,999 a month for 3 to 5 broad queries: prompt testing, publisher placements and a report. The page their ad lands on sells the same three tiers at the same three prices and says “We do not control AI model training or guarantee specific mentions, citations or placements.” That page is noindex, nofollow. The refund policy has twelve sections and none covers this service.; presswhizz.com ad landing (noindex, nofollow, absent from the sitemap) and presswhizz.com/refund-policy/, raw HTML read 2026-09-28

Source: presswhizz-teardown-2026-09-28. Both pages send "Get started" to the same signup. The limit is stated on the page the ad buys; the outcome is promised on the page search engines and AI crawlers index, and a buyer arriving there does not see the limit.

Oxsome (AI Search Optimization)

Could not verify

What the marketing says: “The result: higher rankings on Google and in AI answers”, in the homepage section on its AI agents.; oxsome.com homepage, raw HTML read 2026-09-29
What the entry price buys: The AI Search Optimization agent comes with the 99 USD a month Website plan: structured data, question-shaped headings and consistent listings, written to the site by the agent. Asked how a buyer would know it is working, the module's own FAQ answers “Ask the assistants yourself.” The pages read show no measurement of AI answers at any tier.; oxsome.com/pricing (meta description: 99 and 499 USD a month) and oxsome.com/services/ai-search (FAQPage), raw HTML read 2026-09-29

Source: oxsome-teardown-2026-09-29. The outcome is stated as a result; the only check offered is asking an assistant once, which is one answer, not a measurement.

BluShark Digital (AI Search)

Could not verify

What the marketing says: “Become the trusted answer”, on the AI Search tile of the homepage of a law-firm SEO agency.; blusharkdigital.com homepage, raw HTML read 2026-09-29
What the entry price buys: No price on the homepage, FAQ, case studies or its paid landing page; every path ends at a consultation. A third-party listing reports a 1,000 USD minimum project and 100 to 149 USD an hour. No page read shows an AI metric, a sample or a method behind the AI Search claim.; blusharkdigital.com pages read 2026-09-29; the price range is from Clutch, a third-party listing, read 2026-09-29

Third-party evidence. This vendor does not publish pricing on its own site, so the purchasable half comes from review roundups rather than from them. Treat it as weaker than a first-party quote. Source: sociosquares-blushark-teardown-2026-09-29. First legal-agency row: the buyer /audit/law-firms and Vendor Check serve.

Madcraft (Answer Engine Optimization)

Could not verify

What the marketing says: “Our clients consistently appear in AI Overviews within weeks”, on its Answer Engine Optimization page.; wearemadcraft.com/our-services/answer-engine-optimization/, Wayback capture 2026-05-26 (the live page answers automated readers with a bot challenge, which we do not bypass)
What the entry price buys: No price on the site; every call to action is a consultation form. A third-party listing reports a 25,000 USD minimum project and 100 to 149 USD an hour. The only verification described is simulated queries across ChatGPT, Perplexity and “Google SGE Labs”, with no count, and none of the 15 case studies on its work page is an AEO engagement.; wearemadcraft.com AI search page, Wayback capture 2026-05-26; the price range is from Clutch, a third-party listing, read 2026-09-29

Third-party evidence. This vendor does not publish pricing on its own site, so the purchasable half comes from review roundups rather than from them. Treat it as weaker than a first-party quote. Source: madcraft-teardown-2026-09-29. A timed outcome with no client, count or window, and a test platform named after a Google surface that no longer carries that name.

How to read the verdicts

Stated where you buy: The limit is printed on the page where you hand over money.
True, but not on the purchase path: Accurate somewhere: a blog post, the docs, a third-party roundup, or an upgrade modal you only see after subscribing. Not where a buyer would look first.
Not what the entry price buys: The marketing names a capability the entry price does not include, and the limit is not stated at the point of sale.
Could not verify: We looked and could not confirm it. Said out loud rather than assumed.

Every vendor gates something, and gating is what tiers are for. Nothing here says a limit is wrong. The only thing being measured is whether you can see it before you pay. Rows were read between 2026-05-26 and 2026-09-29, and each row carries its own dates. If a vendor changes a page and a row goes stale, tell us and we will re-read it and update its date.

The full tools directory Our pricing, engines named per tier How we measure