AI visibility checker

Summary#

The Free AI Search Visibility Checker fetches a page along with its robots.txt and llms.txt, scores eleven weighted signals that answer engines rely on, and returns a 0-100 readiness score with a ranked list of fixes. It runs at /free-tools/ai-visibility-checker with no account. Use it to find out whether ChatGPT, Perplexity, Gemini and Google AI Overviews are able to read and quote a page at all.

Purpose#

When someone asks an assistant about your category, the model either recalls your brand from training or retrieves a page about you in the moment. Both paths fail for the same boring reasons: an AI crawler is blocked in robots.txt, the page carries no structured data, or the content is too vague to quote.

Those causes are all visible in your HTML. This tool checks them in one pass so you stop guessing why the assistants name a competitor instead of you.

Overview#

The checker makes three fetches in parallel: the URL you entered, /robots.txt at the same origin, and /llms.txt at the same origin. It then evaluates eleven checks, each carrying a fixed weight, and turns the weights you passed into a percentage.

Every check reports a specific finding rather than a pass or fail alone. If crawlers are blocked, the detail names which ones. If your title is the wrong length, the detail gives the character count.

This is a readiness measure. It tells you whether an engine can use your page. It does not query any model and does not tell you whether an engine did name you. That measurement is Visibility Overview and Prompt Tracking inside the app.

Benefits#

• Finds a blocked AI crawler in seconds, which is the single most common reason a site is absent from AI answers.
• Shows the exact weight behind each signal, so you know which fix is worth most.
• Gives a ranked fix list rather than a bare score.
• Works on any public URL, including a competitor's, with no setup.
• Needs no account, no card and no credits.

Use Cases#

• Pre-launch check. Run it on a new page before you publish and clear the failed checks while the page is still easy to edit.
• Diagnosing absence. An assistant recommends three rivals and never you. Run this first, because a Disallow: / for GPTBot explains it outright.
• Competitor teardown. Run it on the rival the assistants keep naming and compare which signals they carry that you do not.
• Post-migration verification. New CMS, new robots.txt. Confirm the AI bots are still allowed.

Requirements#

A public URL that returns HTML over HTTP or HTTPS. Nothing else. Private, loopback and internal addresses are refused, and a page behind a login cannot be read.

Permissions#

Anyone. No account, no plan, no role. There is one shared protection: an hourly fair-use limit counted per IP address across all the free tools. Normal use never reaches it.

Cost#

Free. Zero credits, and no account is charged because none is involved.

Navigation Path#

metricvaultai.com → Free tools → AI Visibility Checker

The page is at /free-tools/ai-visibility-checker. It is also the first card on the /free-tools index and the first item in the site navigation's Free tools dropdown.

Inputs#

The form's helper line reads No signup · No credit card · Tests the live page, robots.txt and llms.txt.

Step-by-Step Guide#

1. Open /free-tools/ai-visibility-checker.
2. Type the page you want checked into the field marked example.com. Use the full path if you care about a specific article.
3. Select Check AI visibility. The button changes to Checking… and the result area shows Reading the page, robots.txt and llms.txt, and scoring AI readiness….
4. Read the score ring first, then work down the Priority fixes list from the top. It is already sorted by weight, so the first entry is the most valuable change you can make.

A run normally completes in a few seconds. Slow origins take longer because the tool waits on your server, not ours.

Reading the Results#

The score ring. A 0-100 number with a tier badge and a plain-language verdict:

Beside the ring, a heading reports <N> of 11 AI-readiness checks passed, and a How to read it strip repeats the bands as 80-100 strong, AI can cite you, 55-79 partly ready and 0-54 at risk.

What AI engines see. Every check, with a tick or a cross, the specific finding and its point value.

The weights total 100. Substantial content is scored at zero weight on purpose: it is reported as information, not as a penalty, because word count alone is a poor proxy for quality.

The bots checked in robots.txt are GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-Web, anthropic-ai, PerplexityBot, Google-Extended, CCBot and Applebot-Extended. A blanket User-agent: * with Disallow: / fails the check on its own.

Priority fixes. The failed checks, numbered and sorted by weight, each with the same detail text shown above. Work top down.

Examples#

Screenshots#

Tips#

• Check an important article, not just the homepage. Assistants cite pages, and a homepage often scores well while the pages you want quoted score poorly.
• Hover the small i marker beside a check name for a sentence on what the signal is and what a good result looks like.
• The crawler check reads robots.txt at the origin, so it is the same answer for every page on the domain. Fix it once.
• Run the same URL after each fix. The score updates immediately because nothing is cached.

Best Practices#

• Clear the 22-point and 20-point checks before anything else. Crawler access and structured data are worth 42 of the 100 points between them.
• Treat 80 as the working target rather than 100. Substantial content scores zero weight, so the highest achievable weighted total is reached without it.
• Re-check after every template change. Structured data and canonical tags are usually template-level, so one deploy can move every page at once.
• Pair this with AI citation audit on the same URL. This tool asks whether engines can read you, that one asks whether they have a reason to name you.

Common Mistakes#

• Assuming the score predicts mentions. It measures readiness. A perfect 100 on an unknown site will not make ChatGPT recommend you tomorrow.
• Blocking AI crawlers by accident. Boilerplate robots.txt snippets that block "AI scrapers" are common and cost 22 points instantly.
• Reading the homepage score as the site score. Each run covers one URL.
• Expecting the result to be saved. Nothing is stored for an anonymous run. Copy anything you need before you close the tab.

Limitations#

• One URL per run. There is no crawl.
• Readiness only. No model is queried, so the tool cannot report whether any assistant actually cites you.
• Server-rendered HTML only. Content injected by JavaScript after load is not seen, so a heavily client-rendered page can under-score.
• Question / FAQ structure in this tool requires FAQPage or Question schema. Question-shaped headings alone do not pass it here, although they do in GEO audit.
• No history, no export, no scheduling and no Get Recommendations panel. Those are in-app features; see Free tools, no account needed.
• A site that blocks our fetch returns an error rather than a score.

Troubleshooting#

FAQs#

Is this really free? Yes. No account, no card, no credits, and no cap other than the shared hourly fair-use limit.

Does a high score mean ChatGPT will recommend me? No, and we will not claim it does. It means the mechanical barriers are gone. Whether an engine names you also depends on your authority and on what else it finds. Tracking actual mentions over time is Visibility Overview.

How is this different from the GEO Audit and the AI Citation Audit? Same fetch, different scoring sets. This tool scores the broadest set of eleven readability signals. GEO audit weights answer-shaped structure more heavily. AI citation audit weights concrete facts, author and date signals. Running all three on one URL is a reasonable use of five minutes.

Do you store the URL I check? An anonymous run is not attached to an account and is not written to any user history. See What data we store for what the platform stores in general.

Can I run this across every page on my site? Not from the free page. Site-wide checking is Site Audit in the app.

What does the paid product add on top of this? Repeat measurement. The app queries the models on a schedule, records whether you were named, compares you with competitors and alerts you when it changes. Start with What you get on Free.