Agent readiness for content creators
A creator is not a website. You are a name spread across platforms an assistant reads in its own way, and you own almost none of that surface. We measure the three places an assistant learns who you are: the hub you own, the profiles the platforms expose, and what the assistants themselves say when asked. Measure, watch, fix, prove: the same loop.
Agent readiness for content creators
02Sample output
Facts declared
yourhub.comDeclared, never inferred
Three sources
A website is one origin, and the scanner reads it over HTTP. A creator is an entity: a hub you control, accounts on platforms nobody can crawl, and a reputation that lives inside the assistants. Each source is read the way it can be read, and every finding lands in one room.
A domain, a newsletter archive, a podcast site. It runs through the same checks as any site, and the checks that only apply to a shop or a publisher are marked not applicable rather than scored as failures. This is the surface a score can live on, because it is the one you control.
Your channel, read through the platform's official interface and never by scraping. YouTube today, from what the channel publishes. What comes back is a set of findings about how a machine reads your profile, not a score.
Share of answer, pointed at a person. Your niche's real questions go to the engines with web search on, and we count whether your hub or your profiles were cited. A second kind of question asks about you by name, and the answer is checked against the facts you declared.
The outcome you do not control
PaidA brand, a fan or a buyer asks an assistant before they open a search box, and gets back a handful of names. Category questions never name you and measure whether you are among them: cited with a link, mentioned with none, or absent. Accuracy questions name you on purpose: who you are, what you cover, how a brand reaches you. Every answer is kept word for word.
No model grades the answer. The platforms it names are compared with the ones you declared, the contact it gives with the one you chose to publish, the sponsors and affiliations it lists with your list. Anything named that you did not declare is reported as an unverified claim.
We hold what you told us, not what is true. A sponsor an assistant invents is reported as not among your declared facts, never as false. That wording is a rule, because a finding about a person is a claim about a third party.
The channel
Each reading names what an agent cannot find and what it does instead. None of it enters a score: a platform's page belongs to the platform, and the number belongs on the surface you own.
The channel's title and handle against your hub and your other profiles. An agent that cannot line them up cannot tell you from a namesake.
Who you are, what you make and for whom, in the first lines of the description. Without them an agent summarises you from video titles.
A business line in the description, checked only if you opted in to publishing one. Without it an agent invents a mailbox or sends the brand elsewhere.
Your own site, linked wherever the platform lets you link it. Without the link an agent never reaches the one surface written for machines.
The channel keywords, present and consistent with the topics you cover. Without them an agent classifies you from noise.
The share of recent uploads that carry captions. Without them an agent reads titles only, and the content itself is invisible.
Timestamps in recent descriptions. Without them an agent cannot point at the part of a video that answers the question.
One title convention across recent uploads. Without it an agent cannot group a series or tell an episode from a one-off.
The autofix
PaidEverything generated is restructured from what you already publish or declared. Nothing is invented, no email is written unless you opted in, and no credential is ever asked for.
An llms.txt with your name, one sentence, your platforms and your media kit. Person and ProfilePage structured data whose sameAs lists only the profiles you declared. A rel="me" link for each of them.
A description block in the order a machine reads: who, what, for whom, hub, contact. A per-video template with chapters and a naming convention. Captions on.
You apply the change on the platform, and the re-read that confirms it runs on the same decaying schedule as any fix. We never host the hub: a hosted link-in-bio is another product, and the one surface you control should stay yours.
What we write for youThe proof
PaidA brand's errand, run against your hub by an agent with no JavaScript and no privileges, every step recorded. The result is a rate rather than a tick, because something with a temperature was asked to do a job. Without a hub the task reports not applicable, never zero.
Your data
What three models say about a named person is personal data, and so is what they say about the rivals you track. That changes the rules, and the rules are part of the product.
Anything read from a platform expires thirty days after it was read and is deleted by a job. What stays is our finding and its verdict, never the platform's fields.
A channel is read from the platform's public API, as any visitor could read it: no login, no credential, no permission granted. Stop watching whenever you like; deleting the creator deletes what was read.
A tracked rival is a public professional identity: a channel, a handle, a hub. Nothing about a rival is read that the rival did not publish.
Our public index measures websites that published a robots.txt. A person has none, so a creators index lists claimed profiles only.
Pricing
A creator is not a site, so the price is not a site's. Declaring profiles is free and unlimited, the first watched profile, on any platform, is free, and each further one is $5.99 a month. A hub is a site, at a site's price. An agency watches a roster the same way: one subscription, one room per creator, one report per client.
One creator, one watched profile.
Every further profile watched. The hub is a site, at $12 a month.
The thesis behind all of this, argued from data: Read the manifesto