Free AI visibility audit

Is your website
visible to AI?

AI assistants are becoming the new front door to the web. Find out whether ChatGPT, Claude, Perplexity and Google AI can access your site in seconds.

Free · No signup required · Results in seconds

How it works

Three steps to your AI visibility score

01

Enter your URL

Paste your website address. No account or credit card required. The check starts immediately.

02

We analyze access

We read robots.txt and llms.txt for declared rules, then send live requests using each crawler's real user-agent string. An agent only counts as allowed if both signals agree, so a site can't claim access on paper while quietly blocking it in practice.

03

Get your report

Get a 0-100 score, an agent-by-agent access map, a score for every page we analyzed and a list of fixes ranked by impact.

Two signals, not one

A site can declare access in robots.txt and still block it in practice. We check both, so "allowed" always means both signals agree.

Declared rules

We read robots.txt and llms.txt, the same files search engines already use.

Live verification

We also send real requests using each crawler's actual user-agent string to see what happens in practice, not just what's declared.

Sample report

What your report includes

A representative example. This is exactly what you receive the moment your check completes.

example.com
Oct 13, 2026

AI Visibility Score

54out of 100
Grade D

Limited visibility. Most AI platforms cannot see your content.

Visibility Signals

robots.txt
llms.txt
live probe
HTTPS

Homepage responded in 3.2 s

Search Engines2/2

search crawlers powering AI-assisted search results

AI Answer Engines2/4

real-time retrieval when users ask AI assistants

Training ConsentSelective · 1/7

Not scored. Blocking model-training crawlers is a licensing choice, so some training crawlers are allowed, others are turned away.

Agent-by-Agent Access · robots.txt + live probe

Robots.txt means the file disallows this crawler. Server means the live request itself failed. Training crawlers show as opted out rather than blocked, and cost no points either way.

Google SearchALLOWED
Bing / CopilotALLOWED
Google Gemini & AI Overviews trainingALLOWED
OpenAI GPTBot (training)OPTED OUT · robots.txt
ChatGPT live browsingALLOWED
OpenAI search crawlerBLOCKED · server
Anthropic ClaudeBot (training)OPTED OUT · robots.txt
Claude live browsingALLOWED
Perplexity crawlerBLOCKED · robots.txt
Common Crawl (used to train many LLMs)OPTED OUT · robots.txt
ByteDance / TikTok AIOPTED OUT · server
Apple Intelligence trainingOPTED OUT · robots.txt
Meta AI trainingOPTED OUT · robots.txt
4/6
Search & Answer Engines Allowed
1/7
Training Crawlers Allowed
found
robots.txt
missing
llms.txt

How the Score Is Calculated

AI Visibility
67/100
Content Structure
53/100
Technical Infrastructure
25/100
+1 structured data bonus · valid schema, quotable types, freshness dates

Based on robots.txt rules, a live homepage request with each crawler's user agent, and 5 analyzed pages. Cached for 60 minutes.

Performance Matrix · 5 pages

Across 2 hosts. Technical Infrastructure is a property of the host, so pages on the same host share it. AI Visibility varies per page because robots.txt rules are path-scoped.

67
Average score
92
Best performing
docs.example.com
43
Needs attention
example.com/support
11
Unique improvements
Excellent 85+Needs work 50-64Poor 0-49
PageScoreAI Vis.ContentTechnical
Homepage
example.com
no quotable schema12d old
62
+3 schema
677825
Documentation
example.com/docs
no schema210d old
48506625
Support
example.com/support
no schemaundated
43671425
Documentation
docs.example.com
TechArticle, BreadcrumbList5d old
92
+10 schema
838875
API Reference
docs.example.com/api
TechArticle140d old
88
+7 schema
838475

Category Averages

AI Visibility
70/100
Content Structure
66/100
Technical Infrastructure
45/100

Key Insights

  • →Technical Infrastructure: 2 of 5 pages score 75 or better.
  • →Content Structure: 2 pages need work before AI can extract them cleanly.
  • →Structured Data: 2 of 5 pages carry schema an answer engine can quote.
  • →Spread: 49 points separate your best and worst page, so a single site-wide fix will not close the gap.

Recommendations · 11

3high impact5quick wins11unique improvements
Unblock AI search and answer crawlers: OpenAI search crawler, Perplexity crawler
high impactmedium effort

AI Visibility · Affects 5 pages

Homepageexample.com
Documentationexample.com/docs
Supportexample.com/support
Documentationdocs.example.com
API Referencedocs.example.com/api
Serve every page over HTTPS
high impacteasy effort

Technical Infrastructure · Affects 3 pages

Homepageexample.com
Documentationexample.com/docs
Supportexample.com/support
Make paragraphs self-contained
high impactmedium effort

Content Structure · Affects 2 pages

Documentationexample.com/docs
Supportexample.com/support
Create an llms.txt file
medium impacteasy effort

Technical Infrastructure · Affects 5 pages

Homepageexample.com
Documentationexample.com/docs
Supportexample.com/support
Documentationdocs.example.com
API Referencedocs.example.com/api
Create an XML sitemap on example.com
medium impacteasy effort

Technical Infrastructure · Affects 3 pages

Homepageexample.com
Documentationexample.com/docs
Supportexample.com/support
Speed up the homepage response (3.2s)
medium impacthard effort

Technical Infrastructure · Affects 3 pages

Homepageexample.com
Documentationexample.com/docs
Supportexample.com/support
Use semantic HTML5 elements
medium impactmedium effort

Content Structure · Affects 1 page

Supportexample.com/support
Enable server-side rendering
medium impacthard effort

Content Structure · Affects 1 page

Supportexample.com/support
Add structured data (JSON-LD)
medium impacteasy effort

Structured Data · Affects 3 pages

Homepageexample.com
Documentationexample.com/docs
Supportexample.com/support
Publish a dateModified signal on every page
medium impacteasy effort

Structured Data · Affects 1 page

Supportexample.com/support
Refresh content older than 90 days
low impactmedium effort

Structured Data · Affects 2 pages

Documentationexample.com/docs
API Referencedocs.example.com/api

Report generated by isvisible.ai · never shared with third parties

robots.txt scanllms.txt check13 AI agents5 pages analyzed2 hosts
0-100 AI visibility score with grade
Agent-by-agent access map for 13 AI crawlers
A score for every analyzed page, across subdomains
robots.txt, sitemap, llms.txt and HTTPS checks
Recommendations ranked by impact and effort
PDF report download

Ready to see where you stand?

Enter your URL above and get this exact report for your site.

Why it matters

AI is already deciding who gets found

Classic SEO gets your site ranked. AI SEO, also called generative engine optimization (GEO), decides whether AI crawlers can read your site well enough to recommend it at all.

900M+

weekly ChatGPT users

OpenAI's weekly active users more than doubled in the past year. Every one of them could be shown a competitor's site instead of yours.

Source: OpenAI, Feb 2026
~50%

of U.S. Google searches now show an AI Overview

Google now answers roughly half of U.S. searches with an AI-generated summary, often before a single blue link loads.

Source: Google
-25%

projected drop in traditional search volume

Gartner projects traditional search engine volume to fall 25% by 2026 as AI chatbots and assistants take over how people find things.

Source: Gartner

Crawlers

The 13 AI crawlers we check

Search Engines

search crawlers powering AI-assisted search results

GooglebotGoogle's main crawler for classic web search indexing.
BingbotMicrosoft's crawler for Bing search and Copilot's grounding.

AI Training Crawlers

crawlers collecting content for model training

Google-ExtendedControls whether Google can use your content to train Gemini and build AI Overviews, separate from regular search indexing.
GPTBotOpenAI's crawler that collects content to train future GPT models.
ClaudeBotAnthropic's crawler that collects content to train Claude models.
CCBotCommon Crawl's crawler. Its open dataset gets reused to train many different companies' models, not just one.
BytespiderByteDance's crawler, used to train its AI models.
Applebot-ExtendedControls whether Apple can use your content to train Apple Intelligence.
meta-externalagentMeta's crawler for training its AI models.

AI Answer Engines

real-time retrieval when users ask AI assistants

ChatGPT-UserFires when someone asks ChatGPT to browse or read a specific page live.
OAI-SearchBotOpenAI's crawler for ChatGPT's search feature, indexing pages it can cite in answers.
Claude-UserFires when someone asks Claude to browse or fetch a page live.
PerplexityBotPerplexity's crawler, indexes pages to cite in its AI-generated answers.

FAQ

Common questions about AI visibility

It shows how easily AI platforms can find, read and quote your site right now, out of 100 points. Three things drive it: whether search and answer engines are allowed to reach your pages, whether your content is structured so an AI can extract it, and whether the technical basics are in place, including robots.txt, sitemap.xml, llms.txt, HTTPS and a fast response.

robots.txt is a text file at the root of your domain that tells crawlers, including search engines and AI bots, which parts of your site they can access. It's the same file that already controls how Google indexes your pages.

llms.txt is a newer file, published at yourdomain.com/llms.txt, that gives AI systems a clean, direct summary of your site's content. It isn't required yet, but publishing one is an easy way to tell AI answer engines what your site is about.

13 crawlers across three categories: search engines like Googlebot and Bingbot, AI training crawlers like GPTBot, ClaudeBot and Common Crawl, and AI answer engines like ChatGPT's live browsing, Perplexity and OpenAI's search crawler.

Yes. isvisible.ai has a public API, and the exact request format is published at isvisible.ai/llms.txt specifically so AI assistants can read it. Ask ChatGPT, Claude or any AI assistant with web access to check a domain using the isvisible.ai API, and it can send the request and read the score back to you.

ChatGPT, Claude, Perplexity and Google AI Overviews are becoming a new way people find and evaluate businesses. If your site blocks these crawlers, your content can't be cited or recommended in AI answers, even if it ranks well in traditional search.

Yes. Enter your URL and get a full report in seconds, no account or credit card required.

AI Visibility is half the score, Content Structure is a quarter, and Technical Infrastructure is the last quarter. Valid schema markup, quotable schema types and freshness dates add up to 10 bonus points on top. If a site has no robots.txt and the live probe can't run either, access can't be confirmed, so those points count at half weight until one of the signals is available.

Yes. We read your sitemap.xml, spread the selection across different sections of the site rather than ten sibling doc pages, and follow the subdomains your homepage links to. Every page gets its own score, so you can see that your homepage is fine while your docs are unreadable. The headline score still describes the domain you entered.

No. Training crawlers like GPTBot, ClaudeBot and Common Crawl are reported as a consent posture, not scored. Blocking them is a licensing decision and it costs you nothing here. Only search and answer engines count toward your score, because those are what decide whether your site can be found and cited today.