Methodology
AI search engines do not crawl and rank the way Google did. They fetch, they read server-rendered HTML, and they cite sources their models recognize. If your site fails any one of the four signal categories, you are invisible to a growing share of your audience.
The four signals
Access
Can AI crawlers reach your pages?
Readability
Is your content legible to a machine?
Off-site presence
Does the wider web know you exist?
Citation
Do AI engines actually recommend you?
Signal categories
Can AI crawlers reach your pages?
We fetch your robots.txt, check for bot-blocking middleware, and verify that major AI crawlers (OAI-SearchBot, PerplexityBot, Claude-SearchBot, Bingbot) are permitted. A site that blocks these agents is invisible regardless of content quality.
What we check
Checked: robots.txt directives, crawler-specific allow rules, server-level rate limiting responses, and meta robots tags on key pages.
Is your content legible to a machine?
We fetch your pages without JavaScript execution and measure how much meaningful text survives. AI engines read server-rendered HTML. A site that renders entirely in the client leaves AI with an empty page.
What we check
Checked: server-rendered text ratio, heading hierarchy, body length, structured data presence (schema.org JSON-LD), and semantic HTML landmark usage.
Does the wider web know you exist?
Entity recognition matters. AI engines cross-reference your domain against public knowledge sources. We check whether your brand appears in Wikidata and structured directories that AI models draw from at training and retrieval time.
What we check
Checked: Wikidata entity presence, domain registration signals, and whether your brand name resolves to a distinct entity in public knowledge graphs.
Do AI engines actually recommend you?
On paid plans, we run real queries across ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews, then match your domain against the answers. For the grounded engines (Perplexity and Google AI Overviews) we read the actual cited source URLs; for the model engines we record whether your brand is named. We track citation and mention share over time.
What we check
Pro and Scale: 25 or 75 tracked prompts per month across all engines. Free grade includes an on-page citation readiness estimate only.
The Playbook
Every fix in your Playbook links to the published research or crawl data that supports it. Priority, estimated effort, and paste-ready snippets included.
See what Pro unlocksMonitoring
AI answers change weekly. A competitor adds schema markup, a model update shifts citation behavior, and your share-of-voice drops without a single change on your end. Monitoring is what makes SnagTrace useful past day one.
01
Baseline snapshot on every signal category.
02
Automated re-check on a best-effort weekly schedule. We store the delta.
03
If your grade moves by more than one letter, we email you the same day.
04
Real queries across all major AI engines on your tracked prompts.
Score: 54/100
Free. No signup. No writing to your site. You get the grade, the top gaps, and the first three fixes in under a minute.
Grade my site free