Eight checks, all free, and seven of them never ask for an email. Each one answers a single question, and each one links to the full scan when the answer turns out to be complicated.
Everything below is a slice of the same engine. If you do not know which question you have yet, run the graded scan first: it raises the ones that apply to your page and ranks them.
Crawler fetch, browser render, diff and grade. An AI Visibility Score, a Prose Parity percentage, the robots.txt rule for all 15 named crawlers, and a ranked fix list.
Each one exists because somebody kept asking that specific question and did not want a whole report to answer it.
The actual text an AI crawler receives from your page with no browser and no JavaScript. Not a grade, just the document.
Answers: is my content even there? AccessWhich of the 15 named AI crawler tokens your file allows or blocks at the site root, plus every line the parser had to skip.
Answers: am I letting them in? StructureWhether you serve one, whether its structure follows the convention, and how many sections and links it lists.
Answers: do I have a map, and is it stale? ReferenceAll 15 tokens from 9 vendors, what each one controls, and the copy-paste robots.txt lines to allow or block each one.
Answers: which bots exist, and what for? BrandWhat a model says about your brand from memory alone, then what it says after reading only your page. The gap is the finding.
Answers: does AI know who we are? Head to headYou and a competitor scanned in the same run, by the same code, with the gap named and the sub-scores that produced it.
Answers: why them and not us? At scaleUp to 8 URLs in one run, every page graded, a combined score, and the fixes ranked by how many pages they touch.
Answers: is it this page or the template? QuestionsThe questions your page could answer, scored by how much of each answer it already carries, put to real models on the record.
Answers: what could we win?The two questions the fetch tools answer, shown on an example page before you run anything. Green survives to the crawler; red does not.
Same-day courier for independent shops
Flat-rate local delivery across the city, booked in under a minute.
empty div
Live parcel tracking on every job, shared with your customer.
empty div
Simple per-parcel pricing, and no monthly fee.
Two lines are in the response and two are written in by script after load, so a non-rendering crawler never receives them.
GPTBot, allowed by Allow: / in the wildcard group
ClaudeBot, allowed by Allow: / in the wildcard group
PerplexityBot, blocked by Disallow: / in its own group
Google-Extended, blocked by Disallow: / in its own group
A real run covers all 15 named tokens and quotes the line that decided each verdict.
Example content, written for this explanation rather than measured from a real site. Open a tool above for your own numbers.
Reach for it: Reach for this before you assume any AI can describe your product. It is the fastest way to catch the single most common AI-visibility failure: content that only appears after JavaScript runs.
What you learn: The readable text in your initial HTML, the way a non-rendering crawler receives it, as an excerpt of up to 4,000 characters, plus the framework you are on and whether the page came back readable, thin, or blank.
Open What GPTBot seesReach for it: Reach for this after you edit robots.txt, or whenever you are not certain you are letting in the crawlers you actually want. Google retired its own tester, and none of the survivors were built for AI bots.
What you learn: For all 15 named AI crawlers, whether your file allows or blocks each one at the site root, the exact rule that decides, and any syntax that would trip a real parser.
Open robots.txt testerReach for it: Reach for this when you need to allow or block a specific AI bot and want the correct product token, not a guess. It is also the reference to cite when someone asks which crawlers exist.
What you learn: Every major AI crawler in one table: who runs it, whether it is for training, for live AI search, or for an on-demand fetch when someone asks an assistant about your page, the user-agent it sends, official docs, and the copy-paste robots.txt lines to control it.
Open AI crawler directoryReach for it: Reach for this to confirm whether your site serves an llms.txt, the emerging file that tells AI systems which pages matter most, and to see how a good one is structured before you write your own.
What you learn: Whether your-site.com/llms.txt exists, and if it does, its title, how many sections and links it lists, its byte size, and a preview of the file itself.
Open llms.txt checkerWhen someone asks an assistant about your market it does not open your site in a browser. It relies on what its crawler fetched, which is your raw HTML, usually without running any JavaScript.
Each tool requests your page or your robots.txt once, with the crawler's own user agent, and never runs JavaScript. That is the whole point: you see what the bot receives, not what a browser paints.
The extraction, the robots parser and the crawler registry come from the same code behind the graded scanner. These are focused slices of it, not a separate, lesser check.
We fetch from our own servers, so a site that verifies crawlers by IP address can behave differently for the real bot. Every result says so plainly. Evidence, not guarantees.
In this order, because the first item is worth more than the rest put together and the last one costs the least.
Free means free, and it stays that way. No trial clock and no report held back until you talk to somebody. Seven of the eight run without an email; prompt tracking asks for one because it spends real model calls and mails you the result link. Every tool here runs the same code the paid product runs, and returns a link anyone can open without an account.
What a plan buys is repetition: the scheduler that re-runs these checks after every deploy, and the history that turns one number into a trend.
Most do not. Crawlers like GPTBot and ClaudeBot fetch your raw HTML and, unlike Google, generally do not run JavaScript. If your important content is injected by JavaScript after the page loads, those crawlers get an empty shell, and so does any AI answer about you. That gap is what the What GPTBot sees tool exposes.
No. robots.txt only decides whether a crawler is permitted to fetch the page. Even a crawler you welcome gets nothing useful if the page needs JavaScript to render, has no structured content, or blocks it at the server or CDN layer. Access is the doorway; readability is the room.
Classic SEO tools check Google, which renders JavaScript and has its own rules. These tools check the AI layer specifically: how much of your content survives in the raw HTML that non-rendering AI crawlers receive, and which named AI bots your rules allow. It is the same idea people call generative engine optimization, or GEO, and answer engine optimization, or AEO.
Yes, all free, and the fetch tools need no signup at all. The two model-powered checks ask real models live, which costs real compute, so the prompt check asks for your email; the brand check treats it as optional. Nothing to install, no credit card. The paid part is optional: if a tool surfaces a problem, we can implement the fix for you and re-scan to prove the gain.
The 15 named crawlers we track across OpenAI (ChatGPT), Anthropic (Claude), Perplexity, Google, Apple, Meta, ByteDance, Amazon, and Common Crawl. The AI crawler directory lists every one with its user agent and how to control it.
No tool can promise that, and we will not pretend otherwise. What these do is remove the reasons AI cannot read or reach you in the first place. Being readable and reachable is the prerequisite for being cited; it is not a ranking trick.
Or run the graded scan and let it tell you which questions you should have been asking.