How the grades work

Methodology

The launch set was verified Oct 15, 2026. The file's own grading key is: A=independently documented (processor-verified, public record, payer-published rates); B=credible self-report with specifics/track record; C=self-reported only or vendor marketing; D=unverified/hype.

The four grades

A — Independently documented. A payment-processor verification, a public record, or rates the payer published. Not a single social post.
B — Credible self-reported. A specific report with a public track record. The number is still one party's account, so the card says self-reported.
C — Self-reported only. One account, or a vendor's own marketing. Labeled self-reported. No independent check on file.
D — Unverified / scam-watch. No usable proof, or the claim matches hype patterns. Kept visible so the pattern is easy to recognize.

What goes on a card

Every card names the grade in words, describes the business in one sentence, and quotes one revenue or pay figure without cutting it off. Startup cost and time to a first dollar stay on the entry page, and only when the source gave them. The earnings disclaimer appears once at the bottom of the page.

What we quote, and what we write

Dollar figures are quoted as the source stated them. Summaries are written in our own words. A secondary roundup can point us at a lead. Grade A in this set was checked against a primary proof: a processor verification, a public contest record, or rates the payer published. Grades B and C that cite a roundup were not all re-checked against a primary document, and the card says so by staying self-reported.

A post on X starts at grade C, labeled self-reported, unless a reviewer records independent proof and upgrades it. A hype claim stays grade D. Reddit posts are graded from the write-up. The launch set has no Reddit posts.

What a grade is not

It is not a recommendation.
It is not an average you can expect to earn.
It is not for sale. Sponsorship does not move it.
It is not updated by the ingestion script. That script writes a review queue only.

Scam-watch

Grade D stays in the directory on purpose. Hype posts are easier to recognize when they sit next to a definition. Flags on other grades, such as "never pay to join", are warnings about the path, not a claim that the published rate is fake.

The source list for the launch set is on the sources page. Questions about a grade can go to hello@agenticpaycheck.com.

Questions

Does a high grade mean the idea is a good bet?

No. The grade measures the proof behind a reported figure. A documented business can still be a poor fit, and a self-reported story can still be true. We do not turn grades into a ranking of what you should do.

Why are some costs unknown?

Because the source did not report them. Filling the gap with an estimate and presenting it as fact would make the directory less trustworthy, so the field stays unknown.

Can a sponsor buy a better grade?

No. Sponsored placements, if they run later, are labeled. They do not change evidence grades, and they do not reorder the grade itself.

Who publishes a grade change?

A person. The ingestion script can only write review candidates. It cannot edit the published directory.

Agentic Paycheck does not guarantee income, revenue, profit, or employment. Dollar figures are quoted from the named sources. Self-reported claims are labeled self-reported. Independently documented claims describe someone else's result, not a forecast for you. When a source did not give a cost, the entry says it was not disclosed. Sponsored placements, if any, are labeled and do not change evidence grades.