Docs
Docs
How it works
How it works
Send us your site
Send us your website URL and your 3 most important services + city. Five minutes.
We run the checker
We run the CiteScope checker (robots.txt AI-crawler access, llms.txt, sitemap, entity schema, FAQ extractability, answer-ready content) plus a manual prompt-sample review, and score you /100.
Your report lands in 72 hours
Your report lands within 72 hours: every check with evidence, what it costs you, and — on the Fix Pack — the actual corrected files, ready to paste or hand to your web person.
The reference
The 11 checks, in three stages
Reach
| robots.txt present | 8 pts |
| No AI crawler fully blocked in robots.txt | 16 pts |
| At least one AI crawler explicitly addressed in robots.txt | 6 pts |
| llms.txt present & substantive (>=200 chars) | 10 pts |
| sitemap.xml present | 6 pts |
Recognize
| Homepage reachable | 6 pts |
| Entity schema (JSON-LD Organization/LocalBusiness or any @type) | 14 pts |
| Single clear H1 naming the business/offer (>=8 chars) | 8 pts |
| Title + meta description present | 8 pts |
Believe
| FAQ schema or substantive FAQ content (answer-extractable) | 10 pts |
| Substantive homepage text (>=300 words for extraction) | 8 pts |
11 checks · weights sum to exactly 100 · verdict bands: 80+ AI-VISIBLE · 50–79 AI-PARTIALLY-VISIBLE · under 50 AI-INVISIBLE / HIGH RISK
The checker measures technical and structural citation-readiness (Reach / Recognize / Believe inputs), not live AI answers: whether ChatGPT actually names a business in a given session varies by model and prompt and is sampled manually in fulfillment, never promised. FAQ “content” detection is a heuristic (FAQ mention + ≥3 question marks) when no FAQPage schema exists; the report labels it as content-found vs. schema-found distinctly. A run whose every fetch failed is flagged invalid — re-run required — and never ships as a 0/100 report.
FAQ
FAQ
What exactly do you check?
11 checks in three stages — Reach (robots.txt, AI-crawler access, llms.txt, sitemap), Recognize (entity schema, title/meta, H1), Believe (FAQ extractability, substantive answer-ready text). The full checker is our delivery engine; the report quotes its evidence verbatim.
Will this guarantee ChatGPT recommends me?
No — and distrust anyone who promises that. We fix the technical and structural blocks that prevent citation. Citation itself depends on the engines, your reviews, and third-party sources, which we flag in the report but cannot fabricate.
How is this different from AgentShelf (Business #74)?
AgentShelf checks whether an AI shopping agent can buy from an e-commerce store (catalog, checkout, UCP). CiteScope checks whether an AI answer engine can recommend a service business. Different buyer, different checks, different fixes.
How fast?
Audit within 72 hours of intake. Fix Pack within 5 business days. Sentinel re-audits monthly.
What format?
PDF-style report (delivered as a document) + a fix folder: robots-snippet.txt, llms.txt, schema.json, faq-and-headline rewrites. Any web person can install them in under an hour; instructions included.
Is this legal?
Yes. We read only public endpoints (robots.txt, llms.txt, sitemap, homepage) — the same files AI crawlers read. No scraping behind logins, no review manipulation, no fake citations.
Find out whether AI can reach, recognize and cite you.
Run the free self-check above, or send us your site and get the full 11-check audit with evidence — delivered in 72 hours or the audit is free.