Skip to content
Waitlist

Can AI even cite your website?

A brand cannot be cited if AI crawlers are not allowed to read the website, if content only appears via JavaScript or if a page hides its answer beneath several paragraphs of introduction. Citability checks every page to see whether AI systems can read it, understand it, attribute it to your brand and use it as a source.

Join the waitlist View the calculation
CITABILITY · TRANSPARENT CALCULATION 48 pages scored · 6 excluded
63 OUT OF 100
Calculated live from the findings
1 · Readable without JavaScript25 / 25
2 · Answer up front11 / 25
3 · Attributed to the brand19 / 25
4 · Verifiable source8 / 25
Search crawlers allowed – no cap on the metric. TOTAL 63

The gate: robots.txt and llms.txt.

BrandZaps first checks your robots.txt against three classes of AI crawler. For every problematic block, BrandZaps shows the matching line to allow access.

SEARCH CRAWLERS · TOP PRIORITY A block makes citations technically impossible

OAI-SearchBot, PerplexityBot, Claude-SearchBot. Almost always a mistake – caps the metric and triggers a warning.

User-agent: OAI-SearchBot
Allow: /
USER AGENTS A block shuts out the person

ChatGPT-User, Claude-User, Perplexity-User. It affects anyone trying to open the page through an AI right now.

TRAINING CRAWLERS A block is a legitimate choice

GPTBot, ClaudeBot, CCBot, Google-Extended. Costs no citations in search.

llms.txt present Shown, but not included in the metric. FOUND

Four levels worth 25 points each.

The measurement rules are shown in plain language in the interface – as rough as they are. A heuristic should not pass itself off as a verdict.

LEVEL 1 · 25 POINTSReadable without JavaScript

The page delivers at least 200 characters of text without JavaScript.

LEVEL 2 · 25 POINTSAnswer up front

Two complete sentences in the first 600 characters and at least one H2.

LEVEL 3 · 25 POINTSAttributed to the brand

Organization markup is present and one of the stored brand spellings appears in the page text.

LEVEL 4 · 25 POINTSVerifiable source

An author or an updated date is present.

The metric responds to the page set.

Remove a page and the metric changes immediately. Legal pages and noindex pages are automatically excluded from the page set – and this is shown.

Page selection is a tree organised by directory. A directory rule also applies to pages added in later crawls.

PAGE TREE ON · OFF · PARTIAL
/services/ON · 12 pages
/guides/PARTIAL · 18 of 26
/careers/OFF
/legal-notice · /privacyAUTOMATICALLY EXCLUDED

Priority from measurement data.

A page on a topic where a competitor is cited in AI answers gets top priority. The same flaw on a page without such a topic is a side note.

ACTION
PAGE
POINTS GAINED
Move the answer into the first two paragraphs
/services/geo-strategy
+25
Add Organization markup
/facts/geo-audit
+25
Set an updated date
/guides/ (14 pages)
+12
A page with two findings on the same level only counts as improved once both are fixed.
IN THE INTERFACETooltips that name the flaw in plain language
IN THE INTERFACEPages with the greatest impact
IN THE INTERFACEHistory and a cross-check against actual citations
From measurement to revision

The findings for each page also appear in the target sitemap and content gap analysis and feed into the briefs. A technical or content flaw becomes a concrete requirement for the page that needs improving.

Target sitemap → Content gap analysis → Content briefs →

AI can only cite what it can reach, understand and attribute.

A website that AI systems can read and cite – with a clear order of what to do first.

Join the waitlist To GEO optimisation