Is your page in the answer?
Type the question your buyers ask. VectorCite asks Claude with live web search, shows you the answer and every source it cited, and tells you where your page stands: cited, found but passed over, or not found. Then it hands you the fix.
Watch a real check
Not a mock-up: this is a stored record of a question we put to Claude about our own agency, replayed step by step.
You ask
A buyer’s question about your page. Here: “what does Digiocular do”.
Claude searches
Claude runs live web searches and reads what it finds: 9 pages here.
Claude answers
It writes the answer and cites 2 of those pages as its sources.
You get the verdict
Cited, found but passed over, or not found, with a readiness score and the fixes.
Asked Claude Haiku 5.5, with live web searchOct 14, 2026
“what does Digiocular do”
Pages Claude’s search read9 of 9
- digiocular.com/
- digiocular.com/portfolio
- en.wikipedia.org/wiki/Digiscoping
- github.com/gyun-jin/DigiOcular
- digiocular.com/blog
- en.wikipedia.org/wiki/DIGIC
- en.wikipedia.org/wiki/Ocular_(disambiguation)
- en.wikipedia.org/wiki/DigiSync
- mayoclinic.org/drugs-supplements/digoxin-oral-route/description/drg-20072646
How Claude’s answer opens
“Digiocular is most likely the digital agency based in Manila, Philippines. It sells AI-driven web and marketing services to clients abroad, and it also builds its own software products.”
One real check, start to finish
We ran our own agency’s homepage against a question its buyers ask. Claude read 15 pages and cited 8. Ours wasn’t one of them, and the record says why.
digiocular.com
Record 0vqNnSPH8
- Checked
- Oct 14, 2026, 12:48 UTC
- Question
- “AI website agency in Manila”
- Search
- Claude Haiku 5.5 retrieved 15 pages: designrush.com, en.wikipedia.org, minimicegroup.com.ph, high6.com, aipro.ph and more.
- Answer
- “Here are the Manila-area options that look most relevant. Most are self-descriptions, so check their portfolios and request quotes before committing.”
- Sources
- 8 cited: high6.com, netizenworks.com, webdesignphilippines.ph, wazile.com, aipro.ph, minimicegroup.com.ph, lkl.ai, clutch.co.
- Your page
- not among the 15 pages retrieved.
- Readiness
- 66/100 across 47 signals.
- First fix
- Direct answer in the lead. 0 of 5 query terms appear in the first paragraph.
Three verdicts, and no flattery
Every check ends in one of three verdicts. Each one tells you something different to do next.
- What usually causes it
- The page answered the question directly, early, in words the answer could quote.
- What to do next
- Keep it current and check the same question again later: answers vary from run to run.
A real one: digiocular.com, cited for “what does Digiocular do”
- What usually causes it
- Usually the answer is buried, the page lacks evidence worth quoting, or the cited pages cover the question more completely.
- What to do next
- The closest fix: move a direct answer to the top and add the evidence the cited pages have and you don't.
- What usually causes it
- The page may not target this question, may be hard for search to reach, or sits outside the handful of pages search returns.
- What to do next
- Check crawler access first, then the question's wording on the page, then the readiness signals that failed.
A real one: digiocular.com, not found for “AI website agency in Manila”
Separately, every result says whether the answer names your brand at all, cited or not, and how early it does.
What we measure, and what we won't claim
A verdict is only worth acting on if you can see how it was reached. Here is everything that goes into one, and the limits we print on every result.
Measured
- The live answer
- Claude answers your buyer’s question with live web search. You see the answer, every source it cited, and the pages its search read but didn’t use.
- Readiness, 47 signals
- The published rubric, in five classes:
- Structure13
- Content11
- Authority8
- E-E-A-T8
- Trust7
- Relevance, as a distance
- Your page’s passages and the cited pages are embedded with Google’s Gemini embeddings, so “closer to what got cited” is measured, not guessed from keywords.
Not claimed
- One answer is one sample. Answers change from run to run; a free check is a strong signal, not a verdict for all time.
- Gemini isn’t measured. Google’s terms don’t allow monitoring its search-grounded answers.
- Other engines need keys. ChatGPT search, Perplexity, Google AI Overviews and Copilot are checked in the console where their keys are set, or with your own.
- No guarantees. Nobody controls what an engine cites. We measure the gap, hand you the fix, and check again.
Free today. Deeper when you need it.
Free page audit
One page and one buyer question: Claude's live answer, its sources, your verdict, the readiness score and the paste-ready fixes. No signup.
Run a free auditThe console
Check pages on every engine that has a key, watch your questions over time, and re-check after you ship a fix to see whether the answer moved.
See plansDone-for-you
The Digiocular team writes and ships the fixes on your site, then re-checks the answers with you.
How it works
Straight answers
What people ask before they trust a verdict.
What's the difference between SEO, AEO and GEO?
SEO optimises for blue-link rankings. AEO, Answer Engine Optimization, optimises for being cited inside the answer an AI assistant writes. GEO, Generative Engine Optimization, is the wider discipline of being retrievable, quotable and trusted by generative engines. AEO is its sharp end: the citation itself.
Which answer engines does VectorCite check?
The free audit asks Claude, Anthropic's model, with live web search, every time, and shows its answer and sources. The console adds ChatGPT search, Perplexity, Google AI Overviews and Copilot wherever their API keys are set, or with your own keys. Google's Gemini API terms forbid monitoring its search-grounded answers, so we don't measure Gemini.
My page ranks on Google. Why is it 'not found'?
Claude's web search is not Google's ranking. For each question it retrieves a handful of pages and cites a few of them, and a page that ranks well can still be missing from that handful. That gap is exactly what the audit shows you.
Is one answer enough to go on?
It's one sample, and answers vary from run to run. Treat a free check as a strong signal rather than a verdict for all time. The console asks each question more than once and tracks it over time.
How is this different from a keyword tool?
Keyword tools count strings. VectorCite shows you the actual answer and its sources, then measures how close your page's passages sit to the pages that were cited, using embeddings. You see why you weren't cited, not only which words you lack.
Can you guarantee a citation?
No. Nobody can guarantee what a generative engine cites. VectorCite measures the gap, hands you the highest-leverage fix, and checks the answer again after you ship it.
Do I have to rewrite my whole site?
No. Every fix is grounded in your existing page. Usually it is schema.org markup and a few targeted rewrites, not a rebuild.
Check a page now
One page, one question, about half a minute. Free, and no signup. The result is yours to keep, share or print.
