What we check. Named, not mysterious.
The scan has 187 registered detectors. Twenty-five evaluators can contribute measured conditions; 162 observations remain unscored. Applicability and captured inputs determine the denominator for each scan. Inspect all criteria, sources and scoring dispositions.
Free category scores and selected findings. Full findings and export files: $10 once.
Can machines reach you at all?
If gate checks fail, almost everything else is noise. Examples of named checks:
gate.crawlable · gate.indexable · gate.sitemap · gate.robots_quality · gate.renders_no_js · hyg.host_consistency · hyg.soft_404 · hyg.not_found
Good looks like: crawlers get real content without a JavaScript-only shell, sitemaps are sane, the site does not soft-404 money pages.
Can AI systems read you when you want them to?
gate.ai_bot_access · gate.bot_reachability · core.ai_bot_matrix · core.train_vs_retrieve · core.llms_txt · amp.llms_full_txt · core.answer_block · amp.citable_facts · amp.passage_structure · amp.snippet_eligible · amp.question_headers
Good looks like: policy matches intent (training vs search bots split when you care), live probes do not die at the edge while robots say allow, pages open with something an answer engine can quote.
We do not monitor live ChatGPT answers. That is a different product category. See readiness vs monitoring.
Classic search still pays the bills.
core.title · core.meta_description · core.h1 · core.structured_data · core.schema_complete · core.content_depth · amp.canonical · amp.open_graph · amp.schema_breadcrumb · amp.schema_article · amp.schema_faq · core.entity_consistency · core.topic_target
Good looks like: every money page has a clear title and topic, schema matches the page type, thin clone pages are rare. When the audit finds a canonical error, the fix guide names the exact change.
When you sell a product or a place.
Stores: amp.schema_product · amp.schema_merchant_fields (shipping and return on offers) · image and weight hygiene · bot access so AI shopping answers can read PDPs. Door: for ecommerce.
Local and restaurants: core.local_schema · phone and address consistency checks · core.hours_markup · core.schema_visible_parity · core.geo_target. Doors: local · restaurants.
Security, accessibility, privacy, weight.
The 48 security, accessibility and privacy detectors remain available. Seven have narrow captured-input scoring contracts; the remainder are unscored observations. A technical percentage does not certify security, accessibility or legal compliance. See the exact measured conditions.
These are surface scans, not pentests and not legal advice. Standards cited in-product and in docs.
Not every viral GEO claim becomes a check.
We do not invent ranking factors. We do not hard-fail you only because llms.txt is missing. We do not sell "guaranteed ChatGPT citations." If a claim cannot be grounded in operator docs or our own primary measurement, it stays out of the graded set.
Story of the method: how it works. How the grade is actually computed, sources and limits included: methodology. Free tools for single jobs: tools hub. To run your free audit now and see the catalog applied to your site, paste your URL.
Run the catalog against your site.
Free category scores and selected findings. Full findings and export files: $10 once.