BlogGuides

Ten AI visibility products, and three of them fetch your pages

Ten products sold into AI visibility were read at their own published pages on 17 August 2026. Three of them examine the customer's own pages, four read a record of requests that already arrived, and six put prompts to models and read the answers back. The crawler half of the category is bundled free or sold as an annual licence, and the answer half is sold monthly by prompt volume.

17 August 2026 16 min read Lantad

This post reads the published pages of ten products sold into AI visibility and sorts them by what each one obtains. Every plan name, price and capability below was read at the vendor's own page on 17 August 2026 and is reported as that page had it on that date, because a competitor's pricing is a claim about the day you read it and nothing after. Lantad is one of the ten and gets the same treatment as the rest, including the parts that are inconvenient for it. Nothing here is a Lantad measurement of any product: this is a reading of what vendors publish about themselves, not a test of whether what they publish is true.

In short

Of ten products sold into AI visibility and read at their own published pages on 17 August 2026, three examine the customer's own pages: Semrush AI Visibility, Screaming Frog SEO Spider and Lantad.
Four of the ten read a record of requests that already reached the site rather than fetching anything themselves: Cloudflare AI Crawl Control, Ahrefs Bot Analytics, Screaming Frog Log File Analyser and Profound's Agent Analytics.
Cloudflare's AI Crawl Control documentation, read on 17 August 2026, states the feature is available on all plans, and Ahrefs states that Bot Analytics is free during beta and available on all paid Ahrefs plans.
Screaming Frog SEO Spider stores the HTTP response HTML and the rendered HTML side by side for 245 euros a year with no URL limit, which is the same comparison Lantad sells and it runs over a whole site rather than one page.
Lantad publishes Starter at 59, Pro at 199 and Business at 399 US dollars a month in core/src/config.ts, and none of them can be bought today, because PRE_REVENUE in that file is true by a founder decision of 13 August 2026.

Product	Fetches your pages	Reads request records	Queries models	Published price, 17 Aug 2026
Profound	No	Yes	Yes	99 and 399 USD a month, billed yearly
Athena	No	No	Yes	Free tier, then 295 USD a month
Peec	No	No	Yes	Not shown on the pricing page
Knowatoa	No	No	Yes	59 and 199 USD a month
Semrush AI Visibility	Yes	No	Yes	99 USD a month per domain, billed annually
Cloudflare AI Crawl Control	No	Yes	No	Available on all plans
Ahrefs Bot Analytics	No	Yes	No	Free during beta, on paid Ahrefs plans
Screaming Frog Log File Analyser	No	Yes	No	Free to 1k log events, then 139 USD a year
Screaming Frog SEO Spider	Yes	No	No	Free to 500 URLs, then 245 euros a year
Lantad	Yes	No	Yes	59, 199 and 399 USD a month, none on sale

Ten products sold into AI visibility, sorted by what each one obtains. Every cell was read at the vendor's own published page on 17 August 2026 and describes that page on that date. Prices are as published and are not normalised for billing period.

What an AI visibility tool can actually obtain

There are only three places evidence about AI visibility can come from, and every product on this list draws on one or more of them. A tool can put prompts to the models and record the answers. It can read a record of requests that already reached your server, which in practice means log files or an edge provider's own view of its traffic. Or it can fetch your pages itself and inspect what comes back. Nothing else is available, because nothing else exists to look at.

The three are not interchangeable and they fail in different ways. Prompt tracking measures an output that moves on its own: the study covered in seven runs per prompt reported that a per-brand visibility rate estimated from a single run carried a standard error of 0.370, and that seven runs were needed to bring it under 0.10, so a tool that asks once a week is reporting noise as movement unless it says how many runs sit behind the number. Request records are exact about what happened and silent about what would happen: they can tell you that an AI crawler fetched forty of your pages last Tuesday, and nothing at all about the pages it did not fetch or the ones it could not read. A fetch is the only one of the three that can answer a question about a page nobody has requested yet, and it is also the weakest evidence about outcomes, because it observes your site rather than any model's behaviour.

That is the split this post sorts on, and it is worth stating before the names, because criteria chosen after the results are how comparisons get written backwards to a conclusion. Lantad's own answer to the third question is set out on our methodology page, and where a product answers one of the other two better than we do, that is said plainly below. The phrase generative engine optimisation covers all three in ordinary usage, which is a large part of why the shelf is confusing to buy from.

Flow: Can AI read and cite my site? to Query models, read answers; Can AI read and cite my site? to Read request records; Can AI read and cite my site? to Fetch the pages; Query models, read answers to What models say today; Read request records to What crawlers already did; Fetch the pages to What a crawler would receive.

The three kinds of evidence an AI visibility product can obtain, and the question each one can answer. A mechanism diagram, not a measurement of any product.

The six products that put prompts to models

This is the bulk of the market and the machine is broadly the same one in each case: a scheduled prompt runner hits several models, and the answers are diffed into visibility, citation and sentiment metrics. Profound's pricing page, read on 17 August 2026 at https://www.tryprofound.com/pricing, describes running structured prompts across AI platforms daily and tracking citations, sentiment, ranking and competitive presence, and lists Starter at 99 US dollars a month and Growth at 399, both billed yearly, with Enterprise on custom pricing. Our own comparison of what that buys against what a crawlability scan buys is on the Profound comparison page.

Athena publishes the unit of the category more plainly than anyone: its pricing page at https://www.athenahq.ai/pricing states "1 credit = 1 AI response", and on 17 August 2026 listed an Essential tier that is free with 25 dollars of credit, Starter at 295 US dollars a month, and custom Enterprise pricing. The same page offers to manage AI crawler access dynamically with a smart robots.txt and llms.txt, which is a product that writes a file on your site rather than one that reads your pages, and those are different things. Knowatoa, read the same day at https://knowatoa.com, prices Starter at 59 US dollars a month and Growth at 199, and says it tracks seven engines including ChatGPT, Claude, Perplexity, AI Mode, AI Overviews, Gemini and Meta AI. Peec named four tiers, Starter, Pro, Advanced and Enterprise, and showed no price against any of them on the page as read, which is worth knowing before you book a call; the Peec comparison page covers the rest. Otterly is in this group too, and its pricing page returned an empty document to our fetcher twice on 17 August 2026, so no figure for it appears in the table above and the Otterly comparison page is the honest place to look instead.

Semrush belongs here as well, at 99 US dollars a month per domain billed annually for its AI Visibility plan, and it is the only product in this group whose plan description also names the customer's own site. The reason to care about the unit rather than the headline price is that prompt count is the real meter: our note on how many prompts a measurement needs explains why a cheap plan with a small prompt allowance can be worse than no measurement, and the per-platform guidance in how to get cited by ChatGPT assumes you already know whether the crawler got in.

Product	Published tiers	Stated unit of pricing
Profound	Starter 99 USD a month, Growth 399, Enterprise custom	Per month, billed yearly
Athena	Essential free with 25 USD credit, Starter 295 USD a month	1 credit = 1 AI response
Knowatoa	Starter 59 USD a month, Growth 199	Per month, 7 engines tracked
Peec	Starter, Pro, Advanced, Enterprise	No price shown on the page
Semrush AI Visibility	99 USD a month per domain	Per domain, billed annually
Otterly	Page returned no content to our fetcher	Not read, so not reported

The prompt-tracking products, as published on each vendor's own pricing page and read on 17 August 2026. Prices are as shown on that date and are not normalised for billing period or prompt allowance.

The four products that read a record of requests

The second group never asks a model anything and never fetches your site. It reads what already happened, either from your log files or from the edge that served the request, and this is the group where the price of the capability is closest to zero. Cloudflare's AI Crawl Control documentation says the feature gives visibility into which AI services access your content and lets you set allow and block policies, track robots.txt compliance and explore monetisation, and it states availability in three words: available on all plans. Ahrefs is nearly as blunt about price. Its Bot Analytics page, at ahrefs.com/bot-analytics, states that the product is free during beta and available on all paid Ahrefs plans, that it runs entirely through a Cloudflare integration server side, and that no JavaScript and no code change is required to set it up.

Screaming Frog's Log File Analyser is the paid option in this group and the most specific about verification. Its page at https://www.screamingfrog.co.uk/log-file-analyser/ describes uploading your log files, verifying search engine and AI bots, identifying crawled URLs and analysing bot behaviour, lists automatic bot verification as a feature and the ability to view IPs spoofing requests, and prices itself at free for 1,000 log events or 139 US dollars a year to remove the limit. Profound's Agent Analytics sits here as well rather than in the group above: it tracks AI-sourced traffic and attribution across your domains through integrations with Google Analytics, Cloudflare and AWS, which is a record of requests that arrived, not a fetch of anything.

Two limits apply to this whole group and neither is a criticism of any product in it. Verification by address matters because a user agent string is a claim rather than an identity, which is why our AI crawler reference lists tokens and published address ranges together. And a record of requests cannot see a system that never made one: the canary study covered in a robots.txt block did not stop twelve chatbots found that 12 of 18 measurable chatbots still returned site content after every bot had been disallowed, with 10 of the 18 fed by a search engine's crawler rather than by an AI vendor's own. Your logs will show you none of that, because the fetch that mattered was somebody else's.

Cloudflare AI Crawl Control All plans Edge view of AI crawler requests, allow and block policies, robots.txt compliance tracking. Documentation states it is available on all plans.
Ahrefs Bot Analytics Free in beta Server side through a Cloudflare integration, no JavaScript required. Stated as free during beta and available on all paid Ahrefs plans.
Screaming Frog Log File Analyser 139 USD a year Your own log files, with automatic bot verification and a view of addresses spoofing requests. Free to 1,000 log events.
Profound Agent Analytics Bundled AI-sourced traffic and attribution across your domains, taken through Google Analytics, Cloudflare and AWS integrations rather than by fetching.

What each request-record product reads and what it costs, as published on the vendor's own page and read on 17 August 2026.

The three products that fetch your pages

Only three of the ten obtain a page and look at it, and one of them is a desktop crawler that predates the whole category. Screaming Frog's SEO Spider lists "Store & View HTML & Rendered HTML" among its features and says it can crawl as a search bot such as Googlebot or Bingbot, as AI crawlers, or with your own custom user agent. Its page at https://www.screamingfrog.co.uk/seo-spider/ prices it at free to 500 URLs, or 245 euros a year to remove the limit and unlock JavaScript rendering. That combination is the honest comparison point for this site, so it is worth being direct about it: storing the response HTML beside the rendered HTML is exactly the evidence prose parity is computed from, Screaming Frog does it across a whole site under one annual licence, and if what you need is that comparison at scale rather than a graded report on a page, it is the better purchase.

Semrush is the second, and it is the only prompt tracker in this list whose plan description names your own site: the AI Visibility plan lists a site audit for AI readiness alongside the mention tracking. The page does not say what that audit fetches or how it renders, so nothing more specific is claimed here than that the line exists on the plan as read on 17 August 2026.

Lantad is the third. It fetches a URL once over plain HTTP and again in a browser and compares the two, which is what what GPTBot sees reports, and it evaluates robots.txt per crawler token rather than against the wildcard alone. The published account of how that behaves across real sites is the crawlability study. The thing to notice about this column is not who is in it but how small it is: eight of the ten products on this list can tell you something about AI visibility without ever requesting a page from you, which means eight of them are silent on the question of whether the page is readable at all.

A product that fetches the page

Whether the fetch returned 200 or an error
What the response body contained before JavaScript
Whether the rendered DOM held more prose than the response
Which crawler tokens robots.txt allows or blocks
Nothing about whether any model has ever cited it

A product that queries models

Whether a brand was named in an answer
Which URLs the answer cited
How that moved week to week, given enough runs
Nothing about pages no answer happened to touch
Nothing about why a page was not retrievable

What each kind of product can and cannot answer about one specific page. A description of scope, not a measurement of any product.

Where Lantad sits on this list, and what it does not do

Lantad is in the fetch column and also queries models, which makes it one of two products here that touch both halves, the other being Semrush. The published plan ladder lives in core/src/config.ts as PLAN_PRICING and reads Starter at 59, Pro at 199 and Business at 399 US dollars a month, with yearly at ten times monthly rather than twelve, and the same figures are shown on the pricing page. Those are settings in a file, chosen by a founder, not findings about anything.

The inconvenient part of that file is the next constant down. PRE_REVENUE is true, recorded there as a founder decision of 13 August 2026 with the reason given in the comment: not selling the product until it is usable. So every price in the Lantad row of the table above is published and none of it is purchasable today, which is a materially different state from every other row and would be dishonest to leave out of a list this site wrote. The crawler registry that the per-token robots.txt evaluation runs against carries fifteen tokens in core/src/bots.ts, which is again a setting, and it is a smaller number than the total population of AI crawlers rather than a claim to have counted them all.

The scoring has a published limit too, and it is the sharpest one. Our own audit of the scale, a robots.txt that blocks every citation crawler still grades B, found a stored fixture that disallows the six AI crawlers capable of citing the site scoring 78.8 out of 100, because access carries 25 points in the weighting and only 20 of them are lost. A buyer comparing tools on this list should weigh that against the alternative of reading the file directly, which is what the robots.txt tester does without a grade attached. The comparison cluster covers the rest of the field one rival at a time, including the PromptAlpha comparison, and those pages rather than this one are where a head to head belongs.

Fetches the page twice and compares Once over plain HTTP and once in a browser, which is the prose parity comparison.
Evaluates robots.txt per crawler token Against the fifteen tokens in the registry in core/src/bots.ts rather than the wildcard alone.
Reads your server logs or edge records It does not. Cloudflare AI Crawl Control and the Screaming Frog Log File Analyser do.
Crawls a whole site under one licence It does not. Screaming Frog SEO Spider stores response and rendered HTML with no URL limit at 245 euros a year.
Can be bought today PRE_REVENUE is true in core/src/config.ts, a founder decision of 13 August 2026.

Lantad's own row, taken from constants in this repository and from a finding already published on this blog. Weights are settings, not measurements.

Which of the ten to pick, given the question you actually have

The sort is easier once the question is stated precisely, and most of the confusion in this category comes from a buyer asking one question and being sold the answer to the other. If you want to know what models say about your brand and how that moves, buy a prompt tracker, ask what prompt allowance the plan carries and how many runs stand behind each reported number, and treat anything under a handful of runs per prompt as directional at best. Nothing in the fetch column answers that question, this product included.

If you want to know whether AI crawlers are reaching your site in practice, start with the record you can get for nothing. Cloudflare's AI Crawl Control is on every plan, Ahrefs Bot Analytics is free during its beta on paid Ahrefs plans, and between them they cover the question for most sites without a purchase. Paying for a crawlability tool before you have looked at the free edge view is buying an inference when you could have had the observation.

If you want to know whether a specific page is readable at all, or whether a redesign has quietly moved your content behind JavaScript, that needs a fetch, and the choice there is between a desktop crawler over the whole site and a graded report on one page. If you also need a file written rather than a file read, note that llms.txt is a publishing decision rather than a measurement, and that no tool on this list can tell you it changed an outcome. What none of the ten can do is tell you what a model will do with your page once it has read it. Each of them measures one input to that and calls it visibility, and a buyer who knows which input is being measured will get more out of any of them than one who does not.

Flow: What do you need to know? to What models say about us; What do you need to know? to Whether crawlers arrive; What do you need to know? to Whether a page is readable; What models say about us to Prompt tracker, check run count; Whether crawlers arrive to Free edge or log view first; Whether a page is readable to Crawler that fetches and renders.

Choosing by the question rather than by the category label. A decision aid built from the sort in this post, not a ranking.

Written by

Lantad

Published 17 August 2026.

A buyer looking for an AI visibility tool is usually asking one of two questions, and the market sells the answers to both under one name. The first question is what ChatGPT, Perplexity or Gemini say today when somebody asks about your category, which is answered by putting prompts to models on a schedule and reading what comes back. The second is whether those systems can retrieve and read your pages at all, which is answered by fetching the pages, or by looking at what the crawlers did when they arrived. The two questions need different evidence, and a product that answers one of them well tells you nothing about the other.

Common questions

Which AI visibility tools actually check whether AI crawlers can read your site?

Of the ten products read at their own pages on 17 August 2026, three obtain your pages: Screaming Frog SEO Spider, the site audit line on the Semrush AI Visibility plan, and Lantad. A further four read a record of requests that already reached your site rather than fetching it, which answers a related but different question: Cloudflare AI Crawl Control, Ahrefs Bot Analytics, Screaming Frog Log File Analyser and Profound's Agent Analytics.

Is there a free tool for checking AI crawler access?

Yes, for the request-record question. Cloudflare's documentation states that AI Crawl Control is available on all plans, and Ahrefs states that Bot Analytics is free during beta and available on all paid Ahrefs plans, both read on 17 August 2026. For the fetch question, Screaming Frog SEO Spider is free to 500 URLs. Free here means the vendor's own published availability on that date, not a judgement about whether the free tier is sufficient.

Why do most AI visibility tools not crawl your website?

Because they are answering a different question. A prompt tracker measures what models output when asked about your category, and that output can be observed without ever requesting a page from you. Six of the ten products here work that way. It is a real measurement of a real thing; it just carries no information about whether a crawler can retrieve and read the page a model would need in order to cite you.

Where does Lantad sit among AI visibility tools?

It fetches a page over plain HTTP and again in a browser, compares the two, and evaluates robots.txt against the fifteen crawler tokens in its registry, and it also queries models. Its published plans are 59, 199 and 399 US dollars a month, and none of them can be bought today because PRE_REVENUE is true in core/src/config.ts by a founder decision of 13 August 2026. For a whole-site raw against rendered comparison, Screaming Frog SEO Spider does more for 245 euros a year.

See what AI can read on your site

Run a free scan and get a graded report of exactly what AI crawlers can and cannot read, with ranked fixes.

Scan your site free How we measure

In short

What an AI visibility tool can actually obtain

The six products that put prompts to models

The four products that read a record of requests

The three products that fetch your pages

Where Lantad sits on this list, and what it does not do

Which of the ten to pick, given the question you actually have

Common questions

See what AI can read on your site

Related

AI visibility

AI crawler

Generative Engine Optimization (GEO)

Prose Parity

How many prompts an AI visibility measurement needs

A robots.txt that blocks every citation crawler still grades B