A
Page basics
Page loads, title tag, meta description, a single H1 and a canonical link.
Read-only checks for AI search
Is your site blocking AI search crawlers?
Paste one public address and get a pass-or-fix list: seven AI crawlers checked by name in robots.txt, plus your title, description and page text. Free, no signup. Fix what it finds.
Checks one public page, robots.txt and llms.txt. It does not measure rankings or AI mentions.
Checks › One page
Readiness checklist
Real result: our Security Desk home page, 12 of 12 passed01Page loadsPASS: HTTP 200
02Title tagPASS: 53 CHARACTERS
03Meta descriptionPASS: 109 CHARACTERS
04Search-type AI crawlers not blockedPASS
05Text in initial HTMLPASS: ABOUT 855 WORDS
A checklist, not a ranking.No prediction of AI mentions.
One public pagerobots.txt and llms.txtRead-only requestsMCP endpoint at /mcp
01 / What it checks
The technical basics, in plain terms.
Each check reports what it found. A pass is a starting point for review, not a result to sell as demand.
B
Indexing signals
No noindex directive, Open Graph tags and JSON-LD structured data types found.
C
Crawler access
Whether robots.txt exists, which AI crawlers it allows or blocks, and whether a sitemap is declared.
D
Content in the HTML
Roughly how much text is in the initial HTML. Pages that render only in JavaScript can look empty to crawlers.
Also reported, but not counted: whether an optional llms.txt file exists. Its benefit is unproven.
02 / Try it
Check one public page.
Enter a public address such as example.com. The check runs on our endpoint and reads public files only.
No check run yet.
03 / AI crawlers
Seven user agents, checked one by one.
Blocking a training crawler does not necessarily block the search crawler of the same company. Each user agent is read separately.
User agentPurpose
GPTBotOpenAI training crawler
OAI-SearchBotChatGPT search index
ChatGPT-UserChatGPT user-initiated browsing
PerplexityBotPerplexity index
ClaudeBotAnthropic crawler
Google-ExtendedGemini/AI training control token
CCBotCommon Crawl
04 / Connect
Use it from your AI tool.
Signal Check is a read-only MCP service. Add the endpoint to a client that supports remote MCP over HTTP.
MCP endpoint
/mcpFetches one public page plus robots.txt and llms.txt and returns a pass or fix checklist. Takes a url.
A static reference of AI crawler user agents and what each is for. No network access.
The principle
A page can allow a crawler without appearing in an answer.
This check reports the public technical evidence, not a ranking or a promise of citations. Keep robots rules and page signals separate from marketing results.
Plans
Free scan first. $19 a month for saved history.
The check above is free on this page. The Security site runs a free scan of four public files and offers a monthly workspace.
Free scan
Checks the homepage, robots.txt, llms.txt and sitemap.xml. No signup. Results appear on the page and are not saved.
$19 per month
Saved history, change comparison and report downloads. Billed monthly until cancelled, through Polar. Manage or cancel in your Polar customer portal. Checkout collects your email. We create your workspace by hand and email you the sign-in, so access is not instant.
05 / Questions
Plain answers.
Does it measure rankings or AI mentions?
No. It does not query AI answer engines or measure citations. It checks one page plus robots.txt and llms.txt.
Why can a page fail to load?
If the page returns anything other than HTTP 200 to this checker, content checks are skipped. Some sites block automated requests.
Is llms.txt required?
No. It is optional, its benefit is unproven, and it is not counted in the passed and total checks.
Does the checklist score my site?
The count is a checklist count, not a ranking or a prediction of AI mentions.
Can it check private pages?
No. It reads public pages only and does not fetch private network addresses.
How many pages are checked?
One page per run, plus robots.txt and llms.txt on the same site.
Check what crawlers can read.
Read-only. One public page at a time.