Quick Start

Submit any public HTTPS URL to our scanner and receive a technical audit report within seconds. No account is required for a basic scan.

Navigate to the Scanner page.
Enter a public HTTPS website URL.
Review your AI Readiness report and prioritized issues.

What we scan

The engine analyses the homepage plus a small set of important pages — up to five in total — along with the files served at the site root (robots.txt, sitemap.xml and llms.txt) and the HTTP response and headers for each request. Pages are chosen by ranking URLs the scan has already seen in your sitemap and homepage navigation; the scanner does not crawl outward, and it does not execute JavaScript, so content injected by client-side rendering is not evaluated.

Five pages is a ceiling rather than a promise: if your server responds slowly, the scan analyses fewer pages instead of timing out, and the report says which pages it skipped and why. This is a sample of your site, not a full crawl — every report states exactly what was and was not fetched, and no finding describes a page that was not scanned.

The overall score covers the site-level checks and your homepage. The other pages are scored and reported separately, so a page-level problem is visible without silently changing a number you may be comparing against a previous scan.

Validation methodology

The audit engine applies 27 deterministic, rule-driven checks. No LLM is involved in scoring — the same site scanned twice produces the same result. Rules are grouped into five pillars:

AI Accessibility — can AI systems reach the site at all?
AI Understanding — can they interpret what the page is about?
AI Trust — does the site publish the identity signals used to judge credibility?
Content & Retrieval — can useful passages be extracted from it?
Technical Quality — supporting hygiene in how the page is served.

Each finding carries the specific observation behind it, the resource that observation came from, and a confidence value. A check that could not run is reported as such rather than counted as a failure.

AI crawler access

Every report includes a per-crawler access matrix covering 17 known AI crawlers, split by what blocking each one actually costs:

Answers & citations (OAI-SearchBot, PerplexityBot, Claude-SearchBot, Googlebot and others) fetch or index pages so an assistant can answer with them and link back. Blocking one removes your site from that assistant entirely, and this is scored.
Model training (GPTBot, ClaudeBot, CCBot, Google-Extended and others) collect content that may be used to train a model. Blocking these is a legitimate editorial choice that does not affect whether assistants can cite you, so it is reported but never counted against your score.

Generated fixes

Where a finding has a concrete artifact behind it, the report generates one from your site's own observed data — a robots.txt block, JSON-LD structured data, an llms.txt drafted from your real navigation, a sitemap.xml, or the missing <head> tags. These are deterministic, not AI-written: every value traces to something this scan saw, or is a clearly marked placeholder for you to fill in. Findings whose fix is a judgement call get written guidance instead of code.

Rate limits

Anonymous users are limited to a small number of scans per day. Creating a free account raises that limit and unlocks 1 project, which is what lets a re-scan tell you what changed. The Pro plan removes the project cap, keeps your full scan history, re-scans a project automatically every week or month, and emails you when a check that was passing starts failing. Every finding and every generated fix file is included on the free plan. There is no public API today; it is planned, and this page will say so when it exists.

Support

Need help? Contact support@aisearchaudit.pro.