Free AI Visibility Checker for Your Website

Check the AI Visibility of Your Website

To check your website's AI visibility, enter the page address in the free checker above. At Layer3Labs, we run SEO and AI-search routines across the sites in our portfolio, and this is the check we built for them. It reads the page, robots.txt, the sitemap and llms.txt. Then it scores six things AI engines need to crawl, parse and quote a page, out of 100 and graded A to F, and lists the fixes biggest first. It is free, needs no sign-up or email, and stores nothing.

When you enter your website in the field at the top of the page and click Check the AI Visibility of Your Website, our server requests four files: the page itself, /robots.txt, the XML sitemap and /llms.txt. It also requests the page once as OpenAI's search crawler (OAI-SearchBot), to spot a firewall that blocks AI crawlers while letting browsers through.

Reading Your AI Visibility Score

Your overall AI visibility score appears as a letter grade from A to F alongside a total out of 100 points. Our grade scale assigns A to scores of 90 and above, B to 80 through 89, C to 70 through 79, D to 60 through 69, E to 50 through 59, and F to anything below 50. If your page contains a noindex directive in its meta robots tag or response header, the tool caps the maximum possible score at 20.

Six category tiles display your points across the technical foundation required for generative engine optimization (GEO). AI crawler access carries 30 points, Readable without JavaScript carries 15 points, Structured data accounts for 15 points, Answer structure awards 20 points, Trust and freshness represents 12 points, and Discoverability provides the remaining 8 points. The weights are our own judgment, not a vendor standard.

Click any score tile to see its checks and how to fix the ones that lost points. Below the tiles, the What to Fix table lists every check, worst first. Red rows marked Fix scored no points, yellow rows marked Improve scored some, and green rows marked Passed scored all of them. Each row says what was found and how to fix it. The AI Crawlers and Your robots.txt table lists 11 crawlers by role and shows whether your robots.txt allows each one on this page. To send the report to a coworker, use Share Report Link or Email to a Coworker. The link re-runs the check when it opens, so nothing is stored. Copy Report puts the score and fix list on your clipboard as text. Share This Report With My AI copies a ready-made prompt with your results, which you can paste into ChatGPT, Claude or Gemini for step-by-step fixes.

AI Crawler Access and Robots.txt Rules

AI engines rely on distinct automated crawlers for search indexation, live user queries, and offline model training. Search crawlers earn 24 of the 30 crawler-access points: 6 for allowing OAI-SearchBot, 6 for Googlebot, and 4 each for Bingbot, Claude-SearchBot and PerplexityBot. The last 6 go to a page with no noindex or nosnippet directive. We do not deduct points for disallowing training crawlers.

OpenAI documents that OAI-SearchBot surfaces websites in ChatGPT search features, and disallowing it removes a site from search answers while leaving navigational links intact. Similarly, Anthropic notes that Claude-SearchBot indexes content for search responses, whereas ClaudeBot collects data for model training. Perplexity documents that PerplexityBot indexes websites for search results without training foundation models, while Perplexity-User handles immediate user requests.

Your robots.txt rules are parsed according to standard web crawler conventions where specific user-agent groups override generic wildcards and longer rule paths take precedence over shorter ones. The checker alerts you if our OAI-SearchBot request receives a 401, 402, 403, 429, or 503 response while normal page requests succeed. Cloudflare made blocking AI crawlers the default for its customers on 1 July 2025, so search crawlers can be turned away unless the site owner allows them.

Text Rendering Without JavaScript

A Vercel study published on 17 December 2024 found that none of the major AI crawlers it tested run JavaScript during crawls, including GPTBot, ClaudeBot, and PerplexityBot. Text added after load does not count.

The Readable without JavaScript category evaluates the volume of readable prose delivered in the initial server response. Pages containing 300 words or more receive the full 15 points, pages with 150 to 299 words receive 9 points, pages with 50 to 149 words get 4 points, and pages with fewer than 50 words receive 0 points. If your initial HTML payload contains almost no visible text, the checker generates an app-shell warning.

Structured Data and Answer Formatting

Structured data awards up to 15 points in the checker. The Structured data category grants 5 points when valid JSON-LD is present and parses cleanly, 5 points for an Organization or Person entity with sameAs reference links, and 5 points for declaring an appropriate content type such as Article, BlogPosting, Product, Service, FAQPage, HowTo, SoftwareApplication, or LocalBusiness.

Google Search Central says structured data should match the visible text, and that AI Overviews need no special schema. Learn how to configure these entities in our guide to schema markup for AI search.

The Answer structure category awards up to 20 points for formatting that makes a page easy to quote. It gives 3 points each for a 10 to 70 character title, a 50 to 170 character meta description, exactly one H1, at least three H2 or H3 subheadings, a list or table, and paragraphs averaging 80 words or fewer. A heading phrased as a question, or FAQPage schema, adds 2.

Trust, Freshness, and Discoverability Signals

The checker scores trust signals across five technical elements. The Trust and freshness category awards 3 points when an author name appears in schema or visible bylines, 3 points for a published or updated timestamp, 2 points for a canonical tag pointing at the same domain, 2 points for HTTPS, and 2 points for an internal link to an About or Contact page.

The Discoverability category awards 4 points when an XML sitemap is reachable, 2 points for complete Open Graph social metadata, and 2 points for a published /llms.txt file starting with an H1 heading. Jeremy Howard published the llms.txt proposal in September 2024, where the H1 heading is the only required section.

Because Google's documentation explicitly clarifies that AI Overviews require no special AI text files, llms.txt is weighted at only 2 points in our scoring model. You can build your own file with our free llms.txt generator or read our breakdown in llms.txt explained. Homepages frequently drop points for author names and dates because brand landing pages rarely feature bylines, so auditing individual resource articles provides a more realistic assessment of your editorial content.

Limits of the Readiness Score

This AI visibility checker measures crawl and formatting readiness rather than live citations, query rankings, or inbound AI referral traffic. An audit evaluates a single page URL at a time, meaning a site's homepage and an in-depth informational guide can produce very different scores.

Firewall probes executed by our server simulate OpenAI's user agent, but enterprise security networks that verify bot IP addresses may refuse our test request even if actual crawler IPs are permitted. A refused probe is a reason to check your firewall's bot settings, not proof of a block.

To monitor whether automated assistants like ChatGPT, Claude, or Perplexity actively quote your brand across specific conversational prompts, consult our review of the best AI visibility tools. To see how alternative testing utilities compare, review our analysis of free AI visibility checkers. Run the free AI visibility checker on your highest-value URLs, export your prioritized fix list, and resolve crawler disallow rules before updating your schema.