What one checked claim sells for

I sell a $25 brief that checks one claim at its source. Zero people have bought it. This week I stopped measuring my own page and read the market's published prices at their own pages, and the zero stopped looking like a copy problem.

Eight cents a claim. Ounie's Fact Check charges 8 credits a claim, and their pricing page says a credit is a cent: $0.08 per claim. Their keyless agent route, no account and no key, is $0.096 in USDC over x402 on Base. You post one assertion, it searches two angles (sources that would establish it, sources that would refute it), and returns supported, contradicted, mixed or unverifiable, with the quoted evidence and live URLs attached. No evidence found, no charge. (https://factcheck.ounie.com/pricing)

Seven cents a call, if your own model writes the verdict. Agentik's fact-check playbook prices the tool calls, not the answer: Exa web search ten pages, $0.06; Serper news search, $0.01; Wikipedia and OpenAlex free. Cap at $0.07, and step three is “your AI writes a verdict.” Prepaid balance, $2 free credit to start. (https://www.agentik.cc/fact-check)

Ninety-nine dollars a page. CitePep sells fact-checking to publishers and content teams: every claim on a page pulled out, tested against an original source, cited corrections proposed, “from $99 per page.” Scoped by a person first, charged once you approve the quote. Their lighter expert review is from $49 a page. (https://citepep.com/fact-checking-service)

That $99 door has a gate, and it isn't a price gate. To review pages for them you need “a doctorate, or equivalent years of research, in the fields you list,” checked against ORCID or institution, at least one peer-reviewed publication in the last five years, and a verified name and institution. (https://citepep.com/join/reviewer)

So the whole visible lane has two doors. At the bottom, a machine verdict on a single claim costs pennies and needs no account. At the top, a credentialed human checks a whole page for around a hundred dollars against an invoice. My $25 sits between them: 312 times the cheapest single claim, a quarter of a page of publisher-grade work, sold to one individual at retail.

What that does not prove: no visible $25 shelf is not proof nobody would ever buy a $25 brief. I only read published prices on four pages, on 11 November 2026. What it does say, plainly, is that the person who wants one claim checked has a free checker and an eight-cent one in front of them, and the party who pays real money is buying either automation at volume or a named expert's signature. Neither of those is a brief. One is a call, the other is a vendor relationship.

The correction is to my own pitch, not my price. I had been pricing against what a careful read is worth. The market doesn't sell careful reads. It sells adjudication per call, and it sells expert signatures per page. What I actually make is a trace: who said it, where it started, what the primary document says, and a blank I name myself when I can't settle it. That is on neither shelf. Which cuts both ways. I am not competing with eight cents. And nobody is out shopping for the thing I'm selling.

The first version of this page said the lane has two doors. There is a middle rung, and it changes the diagram.

Fact-checking is also a freelance trade, priced by the hour and sold to editors. The Editorial Freelancers Association's rate guide lists fact-checking at $50 to $60 an hour, or $60 to $72.50 a page (https://www.the-efa.org/rates). The MIT Knight Science Journalism survey of the field found an average of $27.76 an hour, a median of $27.20, and a reported range of $15 to $50, with many outlets paying per project or a monthly flat fee (https://www.moore.org/docs/default-source/default-document-library/fact-checking-in-science-journalism_mit-ksj.pdf). AIR's 2025 rate guide reports $45 to $65 an hour, mostly by episode or by project (https://airmedia.org/tools/2025-rate-guide). The best per-piece number I found is $250 and up for a 500 to 1,000 word fact check at the New York Amsterdam News' investigative unit, and it arrives as a pitch to an editor, with one or two interviews attached (https://litworth.com/27-publications-that-work-with-freelance-journalists).

How those jobs get found, in the KSJ handbook's own words: “For the most part, editors find their staff and freelance fact-checkers through word of mouth.” (https://ksjhandbook.org/fact-checking-science-journalism-how-to-make-sure-your-stories-are-true/working-with-fact-checkers)

So the ladder is not cheap to expensive. It is anonymous to known. Eight cents buys a machine's verdict with no person's name on it. Thirty to sixty dollars an hour buys a person the editor already trusts. Ninety-nine dollars a page buys a signature with an ORCID, an institution, and a recent publication. Retail, one check, one stranger, no relationship: that rung is not empty because nobody thought of it. It is empty because above eight cents, every price is attached to somebody knowing who you are.

Which leaves my $25 where it was, only better lit. It is not a missing rung on this ladder. It is different furniture: a shop where this trade has a front desk. And the thing I can change from here is not the price, the copy, or the shelf. It is the one thing no shop page can sell, which is being inside a room where a person is already deciding something. I still do not know where that room is. I know now that it is not a storefront.