VSNARY Emmanuel Orta
In one sentence
CrawlCheck is built by Emmanuel Orta (creator name VSNARY) for VSNARY as the verification layer AI agents read before they act on a website; the local service sites he runs are measured by the same layer, every guide and findings post on this site is written by him from measurements that can be repeated, the pieces that document his own mistakes are listed on purpose, and this page lists all of them with the dates they were published and updated.
I build CrawlCheck, the verification layer for the agentic web, and I write everything published here. VSNARY is my creator name; Emmanuel Orta is my birth name. The sites I run are measured by the same layer every agent reads before it acts on them.
Outside this site I keep a reptile and invertebrate collection and publish photographs of it to Wikimedia Commons under User:Vsnary, each under its own free licence: CC BY 4.0, CC BY-SA 4.0 or CC0. On 9 September 2026 I published the first photograph of a Gonatodes daudini hatchling recorded on Wikimedia Commons, Wikidata, iNaturalist or GBIF — a Critically Endangered gecko. It is now the species’ juvenile image on Wikidata and in the English and Cebuano Wikipedia articles.
.jpg){kind=link}
@id, declared on emmanuelorta.com and reused here rather than re-declared. The profiles on the left are the node’s sameAs; the entities on the right each carry their own identifier on their own site; the files underneath publish the same node. The graph is drawn from the same constants as the page’s JSON-LD, so it cannot disagree with it.How the writing is done
Three rules, applied to every piece on this page. The two figures are computed from the post registry on every load.
Measure
Every piece starts as a measurement the verification layer took: access logs, crawler identities, machine files, structured data, signed answers. Nothing is published that cannot be measured again.
Date it or compute it
A number is either recomputed from the record when the page is opened, or dated to the day it was read. The figures on this page are the first kind.
Correct in place
Where a piece got something wrong, the correction is published in the piece, dated, rather than edited out. The 21 pieces about our own mistakes are listed below on purpose.
Pieces published per week, 2026-08-13 to today — computed from the post registry on every load
What the writing keeps coming back to — glossary terms named by the most pieces
Latest
The six most recent pieces, whatever kind they are.
What I operate
The 4 entities on the right of the graph above, each with the identifier it answers to. CrawlCheck is one of them; my own sites are the test subjects in the guides, so every number can be reproduced by anyone who asks for the domain’s signed answer or runs the free audit. No post names a scanned third-party domain before its owner has been contacted. The full register, re-checked hourly, is at emmanuelorta.com.
Where I got it wrong
21 pieces document a mistake of ours, what it cost, and the rule that stops it recurring. A layer that signs verdicts about other people’s sites owes its readers the same standard.
65 guides
Long-form, each one built around a measurement you can repeat. Newest first, by month.
103 findings
Newest first, by month. Corrections are included here as well as above.
The same person, on other sites
These 17 profiles are the left side of the graph above: the Person node’s sameAs on this site and on emmanuelorta.com, which share one @id. If they ever disagree, the one on emmanuelorta.com is canonical.
emmanuelorta.comGitHubLinkedInWikimedia CommonsFacebookInstagramTikTokcoldblooded.comSubstackQuoraProduct HuntCrunchbaseORCIDThreadsYouTubeMediumGravatar