Buyer's guide
Prefer's free Answer Readiness Scorer grades a page on ten checks with no signup. 8 AEO checkers compared on what they grade, crawler access, schema and cost.
Also inالعربيةPortuguês (Brasil)
The short answer
Prefer's Answer Readiness Scorer is the AEO checker to run first on a page: ten named checks, a pass, partial or fail with a fix on each, a score out of 100, no signup, and nothing uploaded. It is our product, so weigh that. For the site, isvisible.ai probes your homepage live with 13 crawler user agents, and Prefer's Crawler Access Checker reads your robots.txt for 14 named crawlers with the rule behind each verdict. RankSurf, Intura and HubSpot's Grader check the brand rather than the page, and each is one run on one day.
Prefer Answer Readiness Scorer Ten checks, a fix per check, no signup, nothing uploaded 
isvisible.ai 13 crawler user agents fetch your homepage, formula printed 
RankSurf AI Visibility Scan Ten ChatGPT prompts, per-prompt detail, ranked fixes, no signup DisclosurePrefer publishes this list and four of the eight tools on it are ours, so we grouped them by what they check rather than ranking ours as best. We read each other tool's own page in a browser on the dates shown and scored what the page states; we did not create accounts or run every tool ourselves. Every other entry says where it beats Prefer.
In short
How we evaluated
Prefer's Answer Readiness Scorer is the AEO checker to run first on a page: ten named checks, a verdict and fix on each, a score out of 100, no signup, nothing uploaded. Run a crawler checker on the site before that: isvisible.ai probes live, Prefer's reads your robots.txt for 14 crawlers. RankSurf, Intura and HubSpot check the brand, not the page. One run is one sample.
A page grader reads the content of one page. A site checker tests whether crawlers can reach you. A brand checker asks the engines about you. We record which each tool does, from its own page, because a brand score cannot tell you which sentence to rewrite and a page score cannot tell you whether anyone read it.
Whether the tool names its checks or prints its formula. Prefer's scorer lists ten checks; isvisible.ai prints an 80/20 split; HubSpot prints five weights. A number with no criteria behind it cannot be reproduced or acted on.
Whether the result is a verdict per check with the rule or the fix behind it, a breakdown per prompt or dimension, or a single score with a paragraph. We record what each page says the result contains.
Whether the tool fetches your live site, reads it to build prompts or check claims, or reads only what you paste. And the price of a run: an account, an email, a card, a daily cap. Every tool here is free in money.
The scores
Each dimension is scored 0 to 5 on the scale below and the overall is the plain average, so you can check every number against the facts on this page. Prefer is our product and is scored on the same scale.
| Tool | Criteria published | Detail per check | Fixes returned | Reads your live site | Free without signup | Overall |
|---|---|---|---|---|---|---|
| Prefer Scorer (our product) | 5.0 | 5.0 | 5.0 | 2.0 | 5.0 | 4.4/5 |
| isvisible.ai | 5.0 | 5.0 | 1.0 | 5.0 | 5.0 | 4.2/5 |
| Prefer Crawler Checker (our product) | 5.0 | 5.0 | 3.0 | 2.0 | 5.0 | 4.0/5 |
| RankSurf | 3.0 | 4.0 | 5.0 | 3.0 | 4.0 | 3.8/5 |
| HubSpot Grader | 3.0 | 2.0 | 1.0 | 1.0 | 4.0 | 2.2/5 |
How each number is set
Scored 23 September 2026 from what each tool's own page states, read in a browser on the dates in the sources list, plus HubSpot's Grader page read 5 and 6 September 2026 and rendered 8 September. Prefer's rows use the same rules as every other row, from our own tool pages. Three tools are unscored: Prefer's two schema generators check the markup you type rather than a page that exists, so 'reads your live site' and 'detail per check' would grade the wrong thing; Intura's page says no credit card but our note does not record whether it takes an email before the result, so its cost cell cannot be placed and we left the row out rather than guess. We also ask ChatGPT for a free AI visibility checker and read every page it cites; isvisible.ai, RankSurf and Intura are all in the library of pages it cited for those prompts, analysed 23 September 2026. That count is not used in the scores.
Three kinds of checker
Read one page and grade it on a fixed list: answer-first opener, question headings, short sentences, sourced claims, a definition, a date. The scorer reads pasted text, so it grades drafts too. The generators check the markup you are about to add, not the page.
Run on the pages your audit says are losing, and on every draft before it ships. Fix the reds first.Test whether the bots behind AI answers can reach you. isvisible.ai fetches your homepage with 13 user agents; Prefer's reads your robots.txt for 14 named crawlers and tells the search crawlers apart from the training ones.
Run one before any content work. A blocked OAI-SearchBot explains absence that no page grader can.Ask the engines about you and grade the answers. RankSurf builds ten prompts from your domain and ends in ranked fixes; Intura checks the accuracy of AI claims against your site; HubSpot returns a weighted score in one run. Prefer's free checker adds a crawler check on your money pages to the report.
Run one to see where you stand, then the checkers above to see why. Re-run before you trust a number; one run is one sample.The ranked list
Grouped by what the tool grades: page graders, then site crawler checkers, then brand checkers that read your site. Prefer leads each group (our product, disclosed); then by score.
Prefer Answer Readiness ScorerThis is our tool, and it checks a page, not a brand. You paste the visible text of a page (or a draft) and it scores ten things: an answer-first opener, question-shaped headings, short scannable sentences, lists and structure, concrete numbers, sourced claims, a plain 'X is Y' definition, a freshness signal, enough depth, and a summary a model can lift. Each check is marked pass, partial or fail with a one-line fix, and the page tells you to fix the reds first because the opener and headings move the number fastest. It runs in your browser with no signup, so nothing you paste is uploaded, and it does not fetch a URL, because a browser cannot read another site's page. Its own FAQ says a high score does not guarantee a citation (tool page, published 16 August 2026).
Prefer FAQ Schema GeneratorA generator with checks built in, so we list it for exactly what it checks and no more. You add question and answer pairs, or paste a whole FAQ in bulk, and before it outputs one FAQPage object it flags a question with no answer or an answer with no question (and drops the half pair), the same question twice, and tags inside a question or answer. A closing script tag inside your text is escaped. It does not read markup already on your page and it does not validate a live URL. Its page is plain about the context: Google stopped showing FAQ rich results in Search from 7 May 2026, so the markup labels a real FAQ, it no longer earns a search feature, and Google says no special markup is needed to appear in AI Overviews or AI Mode. Free, no signup, runs in your browser (tool page, published 13 September 2026).
Prefer Schema Markup GeneratorThe weakest checker on this list, and we say so because it is ours. It builds six schema types and its check is structural: it emits valid schema.org JSON-LD by construction and leaves out any field you did not fill rather than emitting an empty value. It does not inspect markup that is already on your page, and its own FAQ says the output is still worth a pass through the Schema Markup Validator on the live page, plus a check by eye that every field matches what the page shows. Its type notes carry the dates that matter for an AEO check: FAQ rich results ended in May 2026 and How-to rich results in September 2023, so both types are for clarity, not a search feature. Free, no signup, runs in your browser (tool page, updated 19 September 2026).
Prefer Crawler Access CheckerThis checks a site, or one path on it, for access. You paste your robots.txt, leave the path as a slash or enter a specific one, and it marks each of 14 crawlers allowed or blocked with the rule behind the verdict: OAI-SearchBot, GPTBot and ChatGPT-User for OpenAI; Claude-SearchBot, ClaudeBot and Claude-User for Anthropic; PerplexityBot and Perplexity-User; Googlebot and Google-Extended; Bingbot; Applebot-Extended; CCBot and Bytespider. It resolves groups and longest-match rules the way the crawlers do, with Allow beating Disallow on a tie. It separates the search crawlers (OAI-SearchBot, Claude-SearchBot, PerplexityBot, Googlebot, Bingbot) from the training crawlers and controls (GPTBot, ClaudeBot, Google-Extended, Applebot-Extended), with the roles re-verified against the operators' own docs on 13 September 2026. It reads what you paste, not your live domain, and its page says user-triggered fetchers may not follow robots.txt at all (tool page, updated 19 September 2026).
isvisible.aiA single-page tool that checks a site for access, not a page for content. It reads robots.txt and llms.txt, then runs a live homepage probe with 13 crawler user agents grouped as search engines (Googlebot, Bingbot), training crawlers (Google-Extended, GPTBot, ClaudeBot, CCBot, Bytespider, Applebot-Extended, meta-externalagent) and answer engines (ChatGPT-User, OAI-SearchBot, Claude-User, PerplexityBot). Its formula is on the page: agent access is worth 80 points and llms.txt 20. Free, no signup, results in seconds, with a sample report dated 11 August 2026 and a public API documented in its own llms.txt so assistants can call it (page read 23 September 2026). The result, as its page describes it, is the score, the grade and the access map.
RankSurf AI Visibility ScanA brand checker that starts from your site. Enter a domain and RankSurf auto-generates ten buying-intent prompts, runs them on ChatGPT with live web search, and grades the result on five dimensions it defines on the page: brand mentions, citation quality, prompt coverage, competitor presence, and position and context. You get a score out of ten, a per-prompt breakdown and a ranked fix list at a permanent per-domain URL. It states its limits plainly: point-in-time, ChatGPT only on the free scan, Perplexity and Google AI Overviews on the paid daily scans. Free, no signup, three scans per IP per day (page read 23 September 2026). It reads your domain to build the prompts; it does not grade the pages themselves.
HubSpot AI Search GraderA brand checker, and the one with the largest vendor behind it. HubSpot's own page calls it a free, one-time check with no account required. It queries three models (GPT-5.4 mini, Perplexity and Gemini) and returns an overall score out of 100 with published weights: Sentiment Results 40 points, Presence Quality 20, Brand Recognition 20, Share of Voice 10, Market Competition 10, plus a written interpretation. The checks behind each dimension are not on the page, and the result is the score and the interpretation rather than the answers or the cited pages (page read 5 and 6 September 2026; rendered check 8 September 2026). Am I Cited's September 2026 test also lists it among the checkers that need no account. HubSpot's paid step is HubSpot AEO at $50/mo for 25 prompts on the same three engines.
Intura Free AI Visibility AuditA brand checker from an AI presence agency (Intura Pte. Ltd., Singapore) that reads your site for one thing the others do not: accuracy. It runs category prompts across five engines, checks mentions, citations and share of voice by funnel stage, and grades each claim the engines make about you against your own site as correct, unverifiable, outdated or incorrect. The score formula is on the page: coverage 40%, accuracy 30%, citation 15%, visibility 15%. It returns a fix list and upsells the agency's full audit and ongoing tracking. The page says no credit card and results in minutes; whether it takes an email before the result is not stated in our note, so we have not scored its cost (page read 23 September 2026).
Side by side
What each tool's own page says it checks, at the checked-on date. Partial means adjacent: the schema tools check the markup you type for one page, Prefer's crawler checker tests one path, RankSurf and Intura read your domain to build prompts or check claims but do not grade the pages. The last column is what the page says a run costs you.
| Tool | Checks a page | Checks a site | Checks a brand | Crawler access | Schema | Free without signup |
|---|---|---|---|---|---|---|
| Prefer Scorer | ✓ | - | - | - | - | Yes |
| Prefer FAQ Schema | ~ | - | - | - | ✓ | Yes |
| Prefer Schema Generator | ~ | - | - | - | ✓ | Yes |
| Prefer Crawler Checker | ~ | ✓ | - | ✓ | - | Yes |
| isvisible.ai | - | ✓ | - | ✓ | - | Yes |
| RankSurf | - | ~ | ✓ | - | - | Yes, 3 a day |
| HubSpot Grader | - | - | ✓ | - | - | One run |
| Intura | - | ~ | ✓ | - | - | No card; email not verified |
What changed in 2026
This list looked different a year ago. These are the changes that moved it.
In 2025 an 'AEO checker' meant a brand check: does ChatGPT name you. By 2026 the useful split is page and site graders on one side (Prefer's scorer, isvisible.ai, crawler checkers) and brand checkers on the other. The brand tools tell you where you stand; only the page and site tools tell you what to change.
Crawl access was named in all 12 AI answers Prefer collected for 'what should be on an AEO checklist' in September 2026, yet none of them named OAI-SearchBot, the crawler ChatGPT search uses. Checkers that separate search crawlers from training crawlers, as Prefer's and isvisible.ai's do, are catching up with the operators' own docs.
isvisible.ai prints an 80/20 split, Intura a 40/30/15/15 formula, HubSpot five weights. A year ago a score was a black box. A printed formula is still not a printed check, which is why 'criteria published' is our first scoring dimension.
Google ended FAQ rich results on 7 May 2026, after limiting them to government and health sites in August 2023, and How-to rich results ended in September 2023. A schema check now grades accuracy of labelling, not eligibility for a feature.
Red flags
Every tool here demos well. These are the questions that separate them once the trial ends.
An 'AEO score' out of 100 with no list of checks behind it cannot be reproduced or acted on. Ask to see the checks. Every tool on this list names at least its dimensions; the ones that name every check and the fix for each are the ones you can work from.
Some checkers take your URL or brand, then ask for an email or an account before showing anything. That is fine when the page says so up front, as RankSurf's cap and HubSpot's one run do. Be wary of a tool that hides the gate until you have typed your details in.
A checker that reports 'ChatGPT blocked' because GPTBot is disallowed is wrong: GPTBot is OpenAI's training crawler, and ChatGPT search uses OAI-SearchBot. OpenAI's own docs say each setting is independent. Pick a checker that names both.
A schema checker that promises FAQ or How-to rich results for adding markup is out of date; Google ended both. Add schema to label what is on the page, not to chase a feature.
A brand checker's score is one draw from an answer that moves between runs. Prefer's September 2026 data put ChatGPT's run-to-run overlap of cited sites at 37%. Page and site checks are repeatable; brand checks are samples.
An AEO checker grades a page or a site for answer-readiness: whether the page states its answer first, uses question-shaped headings, backs its claims, carries accurate schema, and can be reached by the crawlers behind AI answers. Prefer’s Answer Readiness Scorer is the page grader on this list, and it is our product, so we say so up front. The rest are the site checkers that test crawler access and the brand checkers that read your site before asking the engines about you. If your question is whether ChatGPT names your brand at all, that is a different tool, and the free AI visibility tools list covers it.
We read each tool’s own page in a browser on 23 September 2026 ( HubSpot‘s Grader on 5 and 6 September, rendered on 8 September), recorded what it says the check grades, returns and costs, and scored those facts on a printed scale. Four of the eight tools are ours, so the list is grouped by what each tool checks rather than ranked with ours on top.
Most tools sold as AEO checking tools are brand checkers. They run prompts against the engines and report whether you were named. That number cannot tell you which sentence to rewrite, and it moves: when Prefer asked 80 questions three times each on 13 September 2026, two ChatGPT runs of the same question shared 37% of the sites cited across both (do AI answers change when asked twice?). A page grader returns the same verdict every time.
A site checker sits in front of both. If OAI-SearchBot cannot reach a page, no answer-first structure will get it cited in ChatGPT search. Prefer’s AEO checklist research found crawl access named in all 12 AI answers we collected, yet none named OAI-SearchBot; the five that named crawlers all listed GPTBot, the training crawler. A checker that tells those two apart is doing something the engines themselves did not.
Prefer’s Answer Readiness Scorer takes the visible text of a page, or a draft, and runs ten checks: an answer-first opener, question-shaped headings, short sentences, lists, concrete numbers, sourced claims, a plain definition, a freshness signal, enough depth, and a summary a model can lift. Each is marked pass, partial or fail with a one-line fix, and the score is out of 100. It reads what you paste, so nothing is uploaded and no signup is needed. Its own page says a high score is necessary but not sufficient: authority off your site still decides the citation. The on-page AEO chapter is the long version of the same checks.
Prefer’s FAQ Schema Generator and Schema Markup Generator are here for exactly what they check. The FAQ tool flags a question with no answer, a repeated question and HTML inside the text before it builds one FAQPage object; the schema tool builds six types that are valid by construction and drops empty fields. Neither reads markup already on a live page; both send you to validator.schema.org for that. Both carry the date that changed what a schema check means: Google ended FAQ rich results on 7 May 2026 and How-to rich results in September 2023 (does schema help AI citations?).
isvisible.ai is the highest-scoring tool here that is not ours. It reads your robots.txt and llms.txt, fetches your homepage with 13 crawler user agents grouped as search engines, training crawlers and answer engines, and returns a score out of 100 with a letter grade and an agent-by-agent map. The formula is printed (agent access 80 points, llms.txt 20) and it needs no signup. Two limits: it probes the homepage only, so a deep Disallow on a money page would not show, and it gives llms.txt 20% of the score, a file no engine has confirmed it reads (does llms.txt help AI citations?).
Prefer’s Crawler Access Checker reads a robots.txt you paste and returns an allowed or blocked verdict for 14 named crawlers, with the exact rule behind each one, for the whole site or a single path. It separates the search crawlers (OAI-SearchBot, Claude-SearchBot, PerplexityBot, Googlebot, Bingbot) from the training crawlers and controls (GPTBot, ClaudeBot, Google-Extended, Applebot-Extended), roles re-verified against the operators’ docs on 13 September 2026. It does not fetch your domain: run isvisible.ai for the live probe and this for the reason (is ChatGPT blocked from my site?).
These three ask the engines about you, and each reads your site on the way in. RankSurf takes a domain, generates ten buying-intent prompts, runs them on ChatGPT with live web search, and returns a score out of ten, a per-prompt breakdown and ranked fixes; no signup, three scans per IP a day, ChatGPT only. Intura runs category prompts across five engines, including Claude and Google AI Overviews, and grades each claim the engines make about you against your own site as correct, unverifiable, outdated or incorrect, with a printed 40/30/15/15 formula; its page says no card, and we could not verify whether it takes an email, so it is unscored. HubSpot’s AI Search Grader queries ChatGPT, Perplexity and Gemini once, no account, and returns a score out of 100 with five printed weights but not the checks behind them.
CitedSpy, Massive, OmniFound and datareport.ai check a brand without reading your site, so they live in the free AI visibility tools list, with Prefer’s own AI Visibility Checker, which adds a crawler check on your money pages to its report.
Site checker first, so you know which engines could reach the page at all. Page grader second, on the pages your AEO audit says are losing; fix the reds. Brand checker third, to see whether the answer moved, and more than once, because one run is one sample. The rest of Prefer’s free tools sit around that loop, all in your browser with no account.
A checker tells you whether a page is ready and whether a bot can reach it. It cannot tell you whether the page you fixed last week changed the answer, because that needs the same prompts tracked over time. Prefer’s plans track your prompts from $24/mo for 30 prompts, with a 7-day free trial; see plans and pricing, or start with the scorer and the crawler checker, which are free.
Every fact about another tool comes from its own page, read in a browser on the date in the sources list; the scores use only those facts and the printed scale, and Prefer’s facts come from our tool pages. We did not test the other tools hands-on; Am I Cited’s September 2026 test is in the sources. We re-read these pages monthly and update the scores when a page changes.
The questions buyers ask next.
Paste the page's visible text into Prefer's free Answer Readiness Scorer. It runs ten checks, from an answer-first opener to sourced claims and a clear definition, marks each pass, partial or fail with a one-line fix, and gives a score out of 100. Above 80 is strong; below 60, start with the opener and the headings. Nothing you paste is uploaded.
Score a page →Eight checks in four groups: crawler access, answer-ready pages, measurement, and presence on other sites. Prefer built the list from 12 AI answers and the operators' crawler docs; the checkers on this page cover the first two groups, and Prefer's measurement covers the third.
The AEO checklist →An AEO audit is a baseline measurement of how AI engines answer your buyer prompts: whether you are cited, your share of voice, and which sources decide each answer. Prefer runs one for free. A checker grades one page or one site on a fixed list. Audit first to find the losing prompts, then check the pages behind them.
What an AEO audit is →Paste your robots.txt into Prefer's Crawler Access Checker and read the OAI-SearchBot row; that is the crawler ChatGPT search uses, and a block there removes you from its search answers. A GPTBot block only affects training. isvisible.ai will also probe your homepage live with both user agents.
Is ChatGPT blocked from my site? →It removes ambiguity; it does not earn a citation. Prefer's schema tools say so on their pages, and Google says no special markup is needed to appear in AI Overviews or AI Mode. Google ended FAQ rich results in May 2026. Label what is on the page accurately, then spend the effort on the answer itself.
Does schema help AI citations? →Questions
An AEO checker grades a page or a site for answer-readiness: whether the page states its answer first, uses question-shaped headings, backs claims with numbers and sources, carries accurate schema, and can be reached by the crawlers behind AI answers. Prefer's Answer Readiness Scorer is a page checker of this kind, with ten named checks and a fix on each. Most tools sold as AEO checkers are AI visibility checkers instead: they ask whether ChatGPT or Gemini names your brand. Both are useful, but they answer different questions, and only the first tells you what to change on the page.
Yes. Prefer's Answer Readiness Scorer is free with no signup: paste a page's text and it returns a score out of 100 across ten checks, each marked pass, partial or fail with a one-line fix, and nothing you paste leaves your browser. For the site, Prefer's Crawler Access Checker and isvisible.ai are both free with no signup; one reads your robots.txt for 14 named crawlers, the other probes your homepage live with 13 user agents. RankSurf's brand scan is free for three runs per IP a day, and HubSpot's Grader is free for one run. Every one of these is a single check; ongoing measurement is the paid product in each case.
An AEO audit measures how AI engines answer your buyer prompts: whether you are cited, your share of voice, and which sources decide each answer. Prefer runs one for free and it ends with targets to work toward. A checker is narrower than an audit: it grades one page or one site on a fixed list, such as Prefer's ten answer-readiness checks or isvisible.ai's 13 crawler probes. Run the checkers on the pages the audit says are losing, in that order: crawler access first, then the page, then measure again.
An AEO checker grades your page or site; an AI visibility checker grades the engines' answers about your brand. Prefer offers both for free: the Answer Readiness Scorer checks one page on ten criteria, and the AI Visibility Checker runs buyer prompts across the engines and records who was named and which pages were cited. RankSurf, Intura and HubSpot's Grader on this list are visibility checkers, which is why they report a brand score rather than a page verdict. If you want the brand tools compared in full, they are in our free AI visibility tools list.
No. Prefer's scorer says so on its own page: a high score makes a page easier to quote, which is necessary but not sufficient, because authority and presence on the third-party sources an engine trusts still decide the citation. In Prefer's September 2026 study, two ChatGPT runs of the same question shared 37% of the sites cited across both, so even a perfect page is one draw from a moving answer. Score the page, then measure the outcome with a visibility check, and repeat the measurement before you call it a win.
The ones that build each engine's search results, not only the training crawlers. Prefer's Crawler Access Checker separates them: OAI-SearchBot for ChatGPT search, Claude-SearchBot for Claude, PerplexityBot for Perplexity, Googlebot for Google AI Overviews and Bingbot for Copilot are the search crawlers; GPTBot, ClaudeBot and Google-Extended control training and grounding. That distinction matters because in Prefer's AEO checklist research, none of the 12 AI answers we collected named OAI-SearchBot, while all five that named crawlers listed GPTBot. isvisible.ai groups its 13 user agents the same way and probes live.
It matters as accurate labelling, not as a route to a search feature. Prefer's schema tools state the dates on their pages: Google stopped showing FAQ rich results on 7 May 2026 and How-to rich results in September 2023, and Google says no special markup is needed to appear in AI Overviews or AI Mode. So a checker that scores you down for missing FAQ schema, or promises a rich result for adding it, is behind. Add schema where it describes what a visitor can see, check it with a validator on the live page, and spend the rest of the effort on the answer itself.
References
How to choose
To fix a page, run Prefer's Answer Readiness Scorer and work down the reds. To find out where you stand, run a brand checker (RankSurf for fixes, HubSpot for a score, or Prefer's free AI Visibility Checker for the answers and cited pages), then come back to the page tools.
If not, do that first. isvisible.ai probes your homepage live; Prefer's Crawler Access Checker reads your robots.txt for 14 crawlers and shows the rule behind each verdict. A blocked OAI-SearchBot or Googlebot explains absence that no page grader can.
Prefer's scorer reads pasted text, so it grades a draft before it ships. isvisible.ai and the brand checkers need a live domain. Score the draft, publish, then probe.
Then a free check will not tell you; that needs the same prompts tracked over time. Prefer's plans track your prompts from $24/mo for 30 prompts with a 7-day free trial. Use the checkers to decide what to change and the tracking to see whether it changed the answer.
Keep going
See how answer engines describe your brand today, and where the openings are to outpace the competition.
5 cookies across three groups. Necessary ones stay on. Change the rest any time from the footer.
Sign-in, security, and remembering this choice. The site does not work without them.
Anonymous page and report usage so we can see what to improve. Nothing is tied to you.
Lets us credit the partner who sent you, so they are paid for the referral, and measure which campaigns bring people here. One cookie, no ad profiling.
